From d21180ab5ea973848d4cdcb896c32400c3d77d38 Mon Sep 17 00:00:00 2001 From: Tom Kraljevic Date: Wed, 8 Apr 2015 14:54:27 -0700 Subject: [PATCH] Renaming of fields in GLM prior1 -> prior beta_constraint -> beta_constraints beta_eps -> beta_epsilon Removed fields in GLM and DL n_folds --- .../packs/examples/Deep Learning Example.flow | 4 +- .../flow/packs/examples/GLM Example.flow | 4 +- .../glm/pyunit_NOFEATURE_benignGLM.py | 3 +- .../glm/pyunit_NOFEATURE_covtypeGLM.py | 6 +-- .../pyunit_NOFEATURE_covtype_getModelGLM.py | 6 +-- ..._NOPASS_perfectSeparation_unbalancedGLM.py | 2 +- .../glm/pyunit_NOPASS_shuffling_largeGLM.py | 6 +-- .../glm/pyunit_link_functions_binomialGLM.py | 2 +- .../glm/pyunit_link_functions_gammaGLM.py | 4 +- .../glm/pyunit_link_functions_gaussianGLM.py | 2 +- .../glm/pyunit_link_functions_poissonGLM.py | 4 +- .../pyunit_perfectSeparation_balancedGLM.py | 2 +- .../glm/pyunit_wide_dataset_largeGLM.py | 2 +- h2o-r/h2o-package/R/glm.R | 53 ++++++++++--------- h2o-r/h2o-package/demo/h2o.glm.R | 2 +- .../glm/runit_GLM_libR_airlines.R | 6 +-- .../glm/runit_GLM_libR_prostate.R | 6 +-- .../glm/runit_GLM_link_functions_binomial.R | 2 +- .../glm/runit_GLM_link_functions_gamma.R | 6 +-- .../glm/runit_GLM_link_functions_gaussian.R | 2 +- .../glm/runit_GLM_link_functions_poisson.R | 4 +- .../runit_GLM_perfectSeparation_balanced.R | 2 +- .../runit_GLM_perfectSeparation_unbalanced.R | 2 +- .../glm/runit_GLM_shuffling_large.R | 6 +-- .../glm/runit_GLM_wide_dataset_large.R | 2 +- .../runit_NOFEATURE_GLMGrid_lambda_search.R | 4 +- .../glm/runit_NOFEATURE_GLM_benign.R | 2 +- .../glm/runit_NOFEATURE_GLM_covtype.R | 6 +-- .../runit_NOFEATURE_GLM_covtype_getModel.R | 6 +-- .../runit_NOFEATURE_GLM_lambda_search_large.R | 6 +-- ...nit_NOFEATURE_GLM_link_functions_tweedie.R | 2 +- .../glm/runit_NOFEATURE_GLM_prostate.R | 2 +- .../tests/testdir_demos/runit_demo_tableau.R | 2 +- .../testdir_demos/runit_demo_tk_cm_roc.R | 2 +- .../runit_NOPASS_Rdoc_glm.R | 4 +- .../runitP_NOPASS_glm2_5_golden.R | 2 +- .../runit_NOPASS_glm2_10_golden.R | 2 +- .../testdir_golden/runit_glm2_11_golden.R | 2 +- .../testdir_golden/runit_glm2_1_golden.R | 2 +- .../testdir_golden/runit_glm2_2_golden.R | 2 +- .../testdir_golden/runit_glm2_3_golden.R | 2 +- .../testdir_golden/runit_glm2_4_golden.R | 2 +- .../runit_glm2_objectiveFun_golden.R | 2 +- .../runit_NOFEATURE_hex_1775_save_load.R | 2 +- ..._NOFEATURE_hex_1799_glm_nfold_parameters.R | 2 +- ...it_NOFEATURE_pub_837_glm_assertion_large.R | 4 +- ...E_pub_874_glm_cv_nonzero_reporting_large.R | 4 +- .../runit_NOPASS_hex_1896_glm_intercepts.R | 6 +-- .../runit_NOPASS_hex_1908_save_load_all.R | 6 +-- .../runit_NOPASS_pub_960_glm_aic.R | 4 +- .../runit_NOPASS_pub_965_binomial_log_pred.R | 2 +- .../runit_hex_1750_strongRules_mem.R | 2 +- .../runit_pub_831_synthetic_strongRules.R | 8 +-- .../runit_pub_838_h2o_perf_message.R | 2 +- .../tests/flow/100KRows2-5Cols.flow | 4 +- h2o-test-integ/tests/flow/100KRows3KCols.flow | 4 +- h2o-test-integ/tests/flow/1MRows3KCols.flow | 4 +- h2o-test-integ/tests/flow/BigCross.flow | 4 +- h2o-test-integ/tests/flow/airlines_all.flow | 4 +- h2o-test-integ/tests/flow/covtype_data.flow | 4 +- h2o-test-integ/tests/flow/prostate.flow | 4 +- 61 files changed, 133 insertions(+), 129 deletions(-) diff --git a/h2o-docs/src/product/flow/packs/examples/Deep Learning Example.flow b/h2o-docs/src/product/flow/packs/examples/Deep Learning Example.flow index 2092c6ed81b2..07202379e7d3 100644 --- a/h2o-docs/src/product/flow/packs/examples/Deep Learning Example.flow +++ b/h2o-docs/src/product/flow/packs/examples/Deep Learning Example.flow @@ -59,7 +59,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-98393399-59c6-4828-97a0-b8d5d458c8f3\",\"training_frame\":\"train.hex\",\"validation_frame\":\"test.hex\",\"dropNA20Cols\":false,\"response_column\":\"C785\",\"n_folds\":0,\"activation\":\"Tanh\",\"hidden\":[50,50],\"epochs\":\"0.1\",\"loss\":\"MeanSquare\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-2362970147619006000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-98393399-59c6-4828-97a0-b8d5d458c8f3\",\"training_frame\":\"train.hex\",\"validation_frame\":\"test.hex\",\"dropNA20Cols\":false,\"response_column\":\"C785\",\"activation\":\"Tanh\",\"hidden\":[50,50],\"epochs\":\"0.1\",\"loss\":\"MeanSquare\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-2362970147619006000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" }, { "type": "md", @@ -82,4 +82,4 @@ "input": "grid inspect \"Training Metrics\", getModel \"deeplearning-98393399-59c6-4828-97a0-b8d5d458c8f3\"" } ] -} \ No newline at end of file +} diff --git a/h2o-docs/src/product/flow/packs/examples/GLM Example.flow b/h2o-docs/src/product/flow/packs/examples/GLM Example.flow index d6475512860e..c7bc585d45d7 100644 --- a/h2o-docs/src/product/flow/packs/examples/GLM Example.flow +++ b/h2o-docs/src/product/flow/packs/examples/GLM Example.flow @@ -39,7 +39,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-072b6da6-fc66-4288-b122-656db078301e\",\"training_frame\":\"abalone1.hex\",\"ignored_columns\":[\"C2\",\"C3\",\"C4\",\"C5\",\"C6\",\"C7\",\"C8\"],\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"ADMM\",\"max_iterations\":-1,\"beta_eps\":0,\"standardize\":false,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.3],\"lambda\":[0.002],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-072b6da6-fc66-4288-b122-656db078301e\",\"training_frame\":\"abalone1.hex\",\"ignored_columns\":[\"C2\",\"C3\",\"C4\",\"C5\",\"C6\",\"C7\",\"C8\"],\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"ADMM\",\"max_iterations\":-1,\"beta_epsilon\":0,\"standardize\":false,\"family\":\"gaussian\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.3],\"lambda\":[0.002],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "md", @@ -58,4 +58,4 @@ "input": "grid inspect \"Coefficient Magnitudes\", getModel \"glm-072b6da6-fc66-4288-b122-656db078301e\"" } ] -} \ No newline at end of file +} diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_benignGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_benignGLM.py index df269ee9f4c6..b845eaa3955d 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_benignGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_benignGLM.py @@ -12,8 +12,7 @@ def benign(ip,port): X = [x for x in range(2,11) if x != Y] #Log.info("Build the model") - model = h2o.glm(y=training_data[Y].asfactor(), x=training_data[X], family="binomial", n_folds=0, alpha=[0], Lambda=[1e-5]) - #model = h2o.glm(y=training_data[Y].asfactor(), x=training_data[X], family="binomial", n_folds=5, alpha=[0], Lambda=[1e-5]) + model = h2o.glm(y=training_data[Y].asfactor(), x=training_data[X], family="binomial", alpha=[0], Lambda=[1e-5]) #Log.info("Check that the columns used in the model are the ones we passed in.") #Log.info("===================Columns passed in: ================") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtypeGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtypeGLM.py index 2f157d2b0558..f17bd53d443b 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtypeGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtypeGLM.py @@ -22,15 +22,15 @@ def covtype(ip,port): #covtype.summary() # L2: alpha = 0, lambda = 0 - covtype_mod1 = h2o.glm(y=covtype[myY], x=covtype[myX], family="binomial", n_folds=0, alpha=[0], Lambda=[0]) + covtype_mod1 = h2o.glm(y=covtype[myY], x=covtype[myX], family="binomial", alpha=[0], Lambda=[0]) covtype_mod1.show() # Elastic: alpha = 0.5, lambda = 1e-4 - covtype_mod2 = h2o.glm(y=covtype[myY], x=covtype[myX], family="binomial", n_folds=0, alpha=[0.5], Lambda=[1e-4]) + covtype_mod2 = h2o.glm(y=covtype[myY], x=covtype[myX], family="binomial", alpha=[0.5], Lambda=[1e-4]) covtype_mod2.show() # L1: alpha = 1, lambda = 1e-4 - covtype_mod3 = h2o.glm(y=covtype[myY], x=covtype[myX], family="binomial", n_folds=0, alpha=[1], Lambda=[1e-4]) + covtype_mod3 = h2o.glm(y=covtype[myY], x=covtype[myX], family="binomial", alpha=[1], Lambda=[1e-4]) covtype_mod3.show() if __name__ == "__main__": diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtype_getModelGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtype_getModelGLM.py index aae859ba3b80..7a679f6d000f 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtype_getModelGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_NOFEATURE_covtype_getModelGLM.py @@ -21,19 +21,19 @@ def covtype_getModel(ip,port): #covtype_data.summary() # L2: alpha = 0, lambda = 0 - covtype_mod1 = h2o.glm(y=covtype[Y], x=covtype[X], family="binomial", n_folds=0, alpha=[0], Lambda=[0]) + covtype_mod1 = h2o.glm(y=covtype[Y], x=covtype[X], family="binomial", alpha=[0], Lambda=[0]) covtype_mod1.show() covtype_mod1 = h2o.getModel(covtype_mod1._key) covtype_mod1.show() # Elastic: alpha = 0.5, lambda = 1e-4 - covtype_mod2 = h2o.glm(y=covtype[Y], x=covtype[X], family="binomial", n_folds=0, alpha=[0.5], Lambda=[1e-4]) + covtype_mod2 = h2o.glm(y=covtype[Y], x=covtype[X], family="binomial", alpha=[0.5], Lambda=[1e-4]) covtype_mod2.show() covtype_mod2 = h2o.getModel(covtype_mod2._key) covtype_mod2.show() # L1: alpha = 1, lambda = 1e-4 - covtype_mod3 = h2o.glm(y=covtype[Y], x=covtype[X], family="binomial", n_folds=0, alpha=[1], Lambda=[1e-4]) + covtype_mod3 = h2o.glm(y=covtype[Y], x=covtype[X], family="binomial", alpha=[1], Lambda=[1e-4]) covtype_mod3.show() covtype_mod3 = h2o.getModel(covtype_mod3._key) covtype_mod3.show() diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_perfectSeparation_unbalancedGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_perfectSeparation_unbalancedGLM.py index 0a201df21c0b..9ca7a3b4d7ca 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_perfectSeparation_unbalancedGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_perfectSeparation_unbalancedGLM.py @@ -10,7 +10,7 @@ def perfectSeparation_unbalanced(ip,port): data = h2o.import_frame(h2o.locate("smalldata/synthetic_perfect_separation/unbalanced.csv")) print("Fit model on dataset.") - model = h2o.glm(x=data[["x1", "x2"]], y=data["y"], family="binomial", lambda_search=True, use_all_factor_levels=True, alpha=[0.5], n_folds=0, Lambda=[0]) + model = h2o.glm(x=data[["x1", "x2"]], y=data["y"], family="binomial", lambda_search=True, use_all_factor_levels=True, alpha=[0.5], Lambda=[0]) print("Extract models' coefficients and assert reasonable values (ie. no greater than 50)") print("Unbalanced dataset") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_shuffling_largeGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_shuffling_largeGLM.py index 48b0b74d53db..e2ada3778f16 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_shuffling_largeGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_NOPASS_shuffling_largeGLM.py @@ -12,13 +12,13 @@ def shuffling_large(ip,port): print("Create model on original Arcene dataset.") - h2o_model = h2o.glm(x=train_data[0:1000], y=train_data[1000], family="binomial", lambda_search=True, alpha=[0.5], n_folds=0, use_all_factor_levels=True) + h2o_model = h2o.glm(x=train_data[0:1000], y=train_data[1000], family="binomial", lambda_search=True, alpha=[0.5], use_all_factor_levels=True) print("Create second model on original Arcene dataset.") - h2o_model_2 = h2o.glm(x=train_data[0:1000], y=train_data[1000], family="binomial", lambda_search=True, alpha=[0.5], n_folds=0, use_all_factor_levels=True) + h2o_model_2 = h2o.glm(x=train_data[0:1000], y=train_data[1000], family="binomial", lambda_search=True, alpha=[0.5], use_all_factor_levels=True) print("Create model on shuffled Arcene dataset.") - h2o_model_s = h2o.glm(x=train_data_shuffled[0:1000], y=train_data_shuffled[1000], family="binomial", lambda_search=True, alpha=[0.5], n_folds=0, use_all_factor_levels=True) + h2o_model_s = h2o.glm(x=train_data_shuffled[0:1000], y=train_data_shuffled[1000], family="binomial", lambda_search=True, alpha=[0.5], use_all_factor_levels=True) print("Assert that number of predictors remaining and their respective coefficients are equal.") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_binomialGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_binomialGLM.py index 6492dc9e32cf..cc99cd9699ec 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_binomialGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_binomialGLM.py @@ -23,7 +23,7 @@ def link_functions_binomial(ip,port): myX = ["ID","AGE","RACE","GLEASON","DCAPS","PSA","VOL","DPROS"] print("Create models with canonical link: LOGIT") - h2o_model = h2o.glm(x=h2o_data[myX], y=h2o_data[myY].asfactor(), family="binomial", link="logit",alpha=[0.5], Lambda=[0], n_folds=0) + h2o_model = h2o.glm(x=h2o_data[myX], y=h2o_data[myY].asfactor(), family="binomial", link="logit",alpha=[0.5], Lambda=[0]) sm_model = sm.GLM(endog=sm_data_response, exog=sm_data_features, family=sm.families.Binomial(sm.families.links.logit)).fit() print("Compare model deviances for link function logit") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gammaGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gammaGLM.py index 6e018393e234..27116060be7c 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gammaGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gammaGLM.py @@ -23,7 +23,7 @@ def link_functions_gamma(ip,port): myX = ["ID","AGE","RACE","GLEASON","DCAPS","PSA","VOL","CAPSULE"] print("Create models with canonical link: INVERSE") - h2o_model_in = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="gamma", link="inverse",alpha=[0.5], Lambda=[0], n_folds=0) + h2o_model_in = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="gamma", link="inverse",alpha=[0.5], Lambda=[0]) sm_model_in = sm.GLM(endog=sm_data_response, exog=sm_data_features, family=sm.families.Gamma(sm.families.links.inverse_power)).fit() print("Compare model deviances for link function inverse") @@ -32,7 +32,7 @@ def link_functions_gamma(ip,port): assert h2o_deviance_in - sm_deviance_in < 0.01, "expected h2o to have an equivalent or better deviance measures" print("Create models with canonical link: LOG") - h2o_model_log = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="gamma", link="log",alpha=[0.5], Lambda=[0], n_folds=0) + h2o_model_log = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="gamma", link="log",alpha=[0.5], Lambda=[0]) sm_model_log = sm.GLM(endog=sm_data_response, exog=sm_data_features, family=sm.families.Gamma(sm.families.links.log)).fit() print("Compare model deviances for link function log") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gaussianGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gaussianGLM.py index 9ab076d1bfd3..908b47af6a8c 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gaussianGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_gaussianGLM.py @@ -23,7 +23,7 @@ def link_functions_gaussian(ip,port): myX = ["ID","AGE","RACE","CAPSULE","DCAPS","PSA","VOL","DPROS"] print("Create models with canonical link: IDENTITY") - h2o_model = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="gaussian", link="identity",alpha=[0.5], Lambda=[0], n_folds=0) + h2o_model = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="gaussian", link="identity",alpha=[0.5], Lambda=[0]) sm_model = sm.GLM(endog=sm_data_response, exog=sm_data_features, family=sm.families.Gaussian(sm.families.links.identity)).fit() print("Compare model deviances for link function identity") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_poissonGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_poissonGLM.py index e2f629af8ae0..cc4ff0f04dac 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_poissonGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_link_functions_poissonGLM.py @@ -22,7 +22,7 @@ def link_functions_poisson(ip,port): myX = ["ID","AGE","RACE","CAPSULE","DCAPS","PSA","VOL","DPROS"] print("Create h2o model with canonical link: LOG") - h2o_model_log = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="poisson", link="log",alpha=[0.5], Lambda=[0], n_folds=0) + h2o_model_log = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="poisson", link="log",alpha=[0.5], Lambda=[0]) print("Create statsmodel model with canonical link: LOG") sm_model_log = sm.GLM(endog=sm_data_response, exog=sm_data_features, family=sm.families.Poisson(sm.families.links.log)).fit() @@ -33,7 +33,7 @@ def link_functions_poisson(ip,port): assert h2o_deviance_log - sm_deviance_log < 0.01, "expected h2o to have an equivalent or better deviance measures" print("Create h2o models with link: IDENTITY") - h2o_model_id = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="poisson", link="identity",alpha=[0.5], Lambda=[0], n_folds=0) + h2o_model_id = h2o.glm(x=h2o_data[myX], y=h2o_data[myY], family="poisson", link="identity",alpha=[0.5], Lambda=[0]) print("Create statsmodel models with link: IDENTITY") sm_model_id = sm.GLM(endog=sm_data_response, exog=sm_data_features, family=sm.families.Poisson(sm.families.links.identity)).fit() diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_perfectSeparation_balancedGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_perfectSeparation_balancedGLM.py index 3c0c5f60744b..e8d8a5bc2219 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_perfectSeparation_balancedGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_perfectSeparation_balancedGLM.py @@ -11,7 +11,7 @@ def perfectSeparation_balanced(ip,port): data = h2o.import_frame(path=h2o.locate("smalldata/synthetic_perfect_separation/balanced.csv")) print("Fit model on dataset") - model = h2o.glm(x=data[["x1", "x2"]], y=data["y"], family="binomial", lambda_search=True, use_all_factor_levels=True, alpha=[0.5], n_folds=0, Lambda=[0]) + model = h2o.glm(x=data[["x1", "x2"]], y=data["y"], family="binomial", lambda_search=True, use_all_factor_levels=True, alpha=[0.5], Lambda=[0]) print("Extract models' coefficients and assert reasonable values (ie. no greater than 50)") print("Balanced dataset") diff --git a/h2o-py/tests/testdir_algos/glm/pyunit_wide_dataset_largeGLM.py b/h2o-py/tests/testdir_algos/glm/pyunit_wide_dataset_largeGLM.py index f23948a93d6f..45ba447d6d4b 100644 --- a/h2o-py/tests/testdir_algos/glm/pyunit_wide_dataset_largeGLM.py +++ b/h2o-py/tests/testdir_algos/glm/pyunit_wide_dataset_largeGLM.py @@ -14,7 +14,7 @@ def wide_dataset_large(ip,port): trainData = h2o.H2OFrame(np.column_stack((trainDataResponse, trainDataFeatures)).tolist()) print("Run model on 3250 columns of Arcene with strong rules off.") - model = h2o.glm(x=trainData[1:3250], y=trainData[0].asfactor(), family="binomial", lambda_search=False, alpha=[1], n_folds=0, use_all_factor_levels=True) + model = h2o.glm(x=trainData[1:3250], y=trainData[0].asfactor(), family="binomial", lambda_search=False, alpha=[1], use_all_factor_levels=True) print("Test model on validation set.") validDataResponse = np.genfromtxt(h2o.locate("smalldata/arcene/arcene_valid_labels.labels"), delimiter=' ') diff --git a/h2o-r/h2o-package/R/glm.R b/h2o-r/h2o-package/R/glm.R index 5522d1536263..ccfdd30e480d 100755 --- a/h2o-r/h2o-package/R/glm.R +++ b/h2o-r/h2o-package/R/glm.R @@ -9,7 +9,7 @@ #' @param destination_key #' @param validation_frame #' @param max_iterations -#' @param beta_eps +#' @param beta_epsilon #' @param score_each_iteration #' @param balance_classes #' @param class_sampling_factors @@ -22,19 +22,19 @@ #' @param tweedie_link_power #' @param alpha #' @param lambda -#' @param prior1 +#' @param prior #' @param lambda_search #' @param nlambdas #' @param lambda_min_ratio #' @param higher_accuracy #' @param use_all_factor_levels -#' @param n_folds +#' @param beta_constraints #' @export h2o.startGLMJob <- function(x, y, training_frame, destination_key, validation_frame, ..., #AUTOGENERATED Params max_iterations = 50, - beta_eps = 0, + beta_epsilon = 0, balance_classes = FALSE, class_sampling_factors, max_after_balance_size = 5.0, @@ -45,21 +45,20 @@ h2o.startGLMJob <- function(x, y, training_frame, destination_key, validation_fr tweedie_variance_power = NaN, tweedie_link_power = NaN, alpha = 0.5, - prior1 = 0.0, + prior = 0.0, lambda = 1e-05, lambda_search = FALSE, nlambdas = -1, lambda_min_ratio = 1.0, use_all_factor_levels = FALSE, - n_folds = 0, - beta_constraint = NULL + beta_constraints = NULL ) { - if (!is.null(beta_constraint)) { - if (!inherits(beta_constraint, "data.frame") && !inherits(beta_constraint, "H2OFrame")) - stop(paste("`beta_constraints` must be an H2OParsedData or R data.frame. Got: ", class(beta_constraint))) - if (inherits(beta_constraint, "data.frame")) { - beta_constraint <- as.h2o(training_frame@conn, beta_constraint) + if (!is.null(beta_constraints)) { + if (!inherits(beta_constraints, "data.frame") && !inherits(beta_constraints, "H2OFrame")) + stop(paste("`beta_constraints` must be an H2OParsedData or R data.frame. Got: ", class(beta_constraints))) + if (inherits(beta_constraints, "data.frame")) { + beta_constraints <- as.h2o(training_frame@conn, beta_constraints) } } dots <- list(...) @@ -96,7 +95,7 @@ h2o.startGLMJob <- function(x, y, training_frame, destination_key, validation_fr parms$x <- args$x_ignore parms$y <- args$y parms$training_frame = training_frame - parms$beta_constraint = beta_constraint + parms$beta_constraints = beta_constraints names(parms) <- lapply(names(parms), function(i) { if (i %in% names(.glm.map)) i <- .glm.map[[i]]; i }) .h2o.startModelJob(training_frame@conn, 'glm', parms, dots$envir) } @@ -119,7 +118,7 @@ h2o.getGLMModel <- function(keys) { h2o.glm <- function(x, y, training_frame, destination_key, validation_frame, #AUTOGENERATED Params max_iterations = 50, - beta_eps = 0, + beta_epsilon = 0, score_each_iteration = FALSE, do_classification = FALSE, balance_classes = FALSE, @@ -132,23 +131,23 @@ h2o.glm <- function(x, y, training_frame, destination_key, validation_frame, tweedie_variance_power = NaN, tweedie_link_power = NaN, alpha = 0.5, - prior1 = 0.0, + prior = 0.0, lambda = 1e-05, lambda_search = FALSE, nlambdas = -1, lambda_min_ratio = 1.0, higher_accuracy = FALSE, use_all_factor_levels = FALSE, - n_folds = 0, - beta_constraint = NULL, + nfolds = 0, + beta_constraints = NULL, ... ) { - if (!is.null(beta_constraint)) { - if (!inherits(beta_constraint, "data.frame") && !inherits(beta_constraint, "H2OFrame")) - stop(paste("`beta_constraints` must be an H2OParsedData or R data.frame. Got: ", class(beta_constraint))) - if (inherits(beta_constraint, "data.frame")) - beta_constraint <- as.h2o(training_frame@conn, beta_constraint) + if (!is.null(beta_constraints)) { + if (!inherits(beta_constraints, "data.frame") && !inherits(beta_constraints, "H2OFrame")) + stop(paste("`beta_constraints` must be an H2OParsedData or R data.frame. Got: ", class(beta_constraints))) + if (inherits(beta_constraints, "data.frame")) + beta_constraints <- as.h2o(training_frame@conn, beta_constraints) } dots <- list(...) @@ -165,7 +164,7 @@ h2o.glm <- function(x, y, training_frame, destination_key, validation_frame, if( missing(x) ) stop("`x` is missing, with no default") if( missing(y) ) stop("`y` is missing, with no default") - if( missing(training_frame) ) stop("`training_frame` is missing, with no default") + if( missing(training_frame) ) stop("`training_frame` is missing, with no default") if (!inherits(training_frame, "H2OFrame")) tryCatch(training_frame <- h2o.getFrame(training_frame), @@ -180,10 +179,16 @@ h2o.glm <- function(x, y, training_frame, destination_key, validation_frame, parms <- as.list(match.call(expand.dots = FALSE)[-1L]) parms$... <- NULL + # For now, accept nfolds in the R interface if it is 0 or 1, since those values really mean do nothing. + # For any other value, error out. + # Expunge nfolds from the message sent to H2O, since H2O doesn't understand it. + if (nfolds > 1) stop("nfolds >1 not supported") + parms$nfolds <- NULL + args <- .verify_dataxy(training_frame, x, y) parms$x <- args$x_ignore parms$y <- args$y - parms$beta_constraint <- beta_constraint + parms$beta_constraints <- beta_constraints names(parms) <- lapply(names(parms), function(i) { if (i %in% names(.glm.map)) i <- .glm.map[[i]]; i }) m <- .h2o.createModel(training_frame@conn, 'glm', parms, dots$envir) m@model$coefficients <- m@model$coefficients_table[,2] diff --git a/h2o-r/h2o-package/demo/h2o.glm.R b/h2o-r/h2o-package/demo/h2o.glm.R index 8baf86fe4c1b..2d2443e8eadf 100644 --- a/h2o-r/h2o-package/demo/h2o.glm.R +++ b/h2o-r/h2o-package/demo/h2o.glm.R @@ -7,7 +7,7 @@ localH2O = h2o.init(ip = "localhost", port = 54321, startH2O = TRUE) prostate.hex = h2o.uploadFile(localH2O, path = system.file("extdata", "prostate.csv", package="h2o"), key = "prostate.hex") summary(prostate.hex) -prostate.glm = h2o.glm(x = c("AGE","RACE","PSA","DCAPS"), y = "CAPSULE", training_frame = prostate.hex, family = "binomial", n_folds = 10, alpha = 0.5) +prostate.glm = h2o.glm(x = c("AGE","RACE","PSA","DCAPS"), y = "CAPSULE", training_frame = prostate.hex, family = "binomial", alpha = 0.5) print(prostate.glm) myLabels = c(prostate.glm@model$x, "Intercept") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_airlines.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_airlines.R index 0d6ebaf2dc39..7c50e42497ee 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_airlines.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_airlines.R @@ -27,18 +27,18 @@ test.LiblineaR.airlines <- function(conn) { Log.info(" family = 'binomial': Logistic Regression\n") Log.info(" lambda = 1/ (cost * params) [3.8e-05]: Shrinkage Parameter\n") Log.info(" alpha = 0.0: Elastic Net Parameter\n") - Log.info("beta_eps= 1E-04: Tolerance of termination criterion\n") + Log.info("beta_epsilon= 1E-04: Tolerance of termination criterion\n") Log.info(" nfolds = 1: No kfold cross-validation\n") h2o.m <- h2o.glm(x = c("DepTime", "ArrTime", "Distance"), #c("fYear","fMonth","fDayofMonth","fDayOfWeek","DepTime","ArrTime","UniqueCarrier","Origin","Dest","Distance"), y = "IsDepDelayed_REC", training_frame = trainhex, family = "binomial", - n_folds = 1, + nfolds = 1, lambda = 1 / (3*100), alpha = 0.0, standardize = TRUE, - beta_eps = 1E-4) + beta_epsilon = 1E-4) h2op <- predict(h2o.m, testhex) h2operf <- h2o.performance(h2o.m, testhex) diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_prostate.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_prostate.R index 43ba9bb6783c..0019e68116f7 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_prostate.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_libR_prostate.R @@ -28,16 +28,16 @@ test.LiblineaR <- function(conn) { Log.info(" family = 'binomial': Logistic Regression\n") Log.info(" lambda = 1/700: Shrinkage Parameter\n") Log.info(" alpha = 0.0: Elastic Net Parameter\n") - Log.info("beta_eps = 1E-02: Tolerance of termination criterion\n") + Log.info("beta_epsilon = E-02: Tolerance of termination criterion\n") Log.info(" nfolds = 1: No kfold cross-validation\n") h2o.m <- h2o.glm(x = c("GLEASON","DPROS","PSA","DCAPS","AGE","RACE","VOL"), y = "CAPSULE", training_frame = trainhex, family = "binomial", - n_folds = 1, + nfolds = 1, lambda = 1/ (7 * 100), #700, alpha = 0.0, - beta_eps = 1E-2) + beta_epsilon = 1E-2) h2op <- predict(h2o.m, testhex) h2opreds <- head(h2op, nrow(h2op)) diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_binomial.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_binomial.R index 9f7b8779c0a6..e2181390186f 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_binomial.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_binomial.R @@ -22,7 +22,7 @@ test.linkFunctions <- function(conn) { R.formula <- (R.data[,"CAPSULE"]~.) print("Create models with canonical link: LOGIT") - model.h2o.binomial.logit <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="binomial", link="logit",alpha=0.5, lambda=0, n_folds=0) + model.h2o.binomial.logit <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="binomial", link="logit",alpha=0.5, lambda=0, nfolds=0) model.R.binomial.logit <- glm(formula=R.formula, data=R.data[,4:10], family=binomial(link=logit), na.action=na.omit) print("Compare model deviances for link function logit") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gamma.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gamma.R index 43004b64b948..4441e5fa1600 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gamma.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gamma.R @@ -23,7 +23,7 @@ test.linkFunctions <- function(conn) { R.formula = (R.data[,"DPROS"]~.) print("Create models with canonical link: INVERSE") - model.h2o.gamma.inverse <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gamma", link="inverse",alpha=0.5, lambda=0, n_folds=0) + model.h2o.gamma.inverse <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gamma", link="inverse",alpha=0.5, lambda=0, nfolds=0) model.R.gamma.inverse <- glm(formula=R.formula, data=R.data[,c(1:5,7:9)], family=Gamma(link=inverse), na.action=na.omit) print("Compare model deviances for link function inverse") @@ -37,7 +37,7 @@ test.linkFunctions <- function(conn) { } print("Create models with link function: LOG") - model.h2o.gamma.log <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gamma", link="log",alpha=0.5, lambda=0, n_folds=0) + model.h2o.gamma.log <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gamma", link="log",alpha=0.5, lambda=0, nfolds=0) model.R.gamma.log <- glm(formula=R.formula, data=R.data[,c(1:5,7:9)], family=Gamma(link=log), na.action=na.omit) print("Compare model deviances for link function log") @@ -51,7 +51,7 @@ test.linkFunctions <- function(conn) { } print("Create models with link: IDENTITY") - model.h2o.gamma.identity <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gamma", link="identity",alpha=0.5, lambda=0, n_folds=0) + model.h2o.gamma.identity <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gamma", link="identity",alpha=0.5, lambda=0, nfolds=0) model.R.gamma.identity <- glm(formula=R.formula, data=R.data[,c(1:5,7:9)], family=Gamma(link=identity), na.action=na.omit) print("Compare model deviances for link function identity") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gaussian.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gaussian.R index 9811e69dfca4..6db7382ca584 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gaussian.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_gaussian.R @@ -24,7 +24,7 @@ test.linkFunctions <- function(conn) { R.formula = (R.data[,"GLEASON"]~.) print("Create models with canonical link: IDENTITY") - model.h2o.gaussian.identity <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gaussian", link="identity",alpha=0.5, lambda=0, n_folds=0) + model.h2o.gaussian.identity <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="gaussian", link="identity",alpha=0.5, lambda=0, nfolds=0) model.R.gaussian.identity <- glm(formula=R.formula, data=R.data[,2:9], family=gaussian(link=identity), na.action=na.omit) print("Compare model deviances for link function identity") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_poisson.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_poisson.R index 018701b23736..7d89c458188e 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_poisson.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_link_functions_poisson.R @@ -22,7 +22,7 @@ test.linkFunctions <- function(conn) { R.formula = (R.data[,"GLEASON"]~.) print("Create models with canonical link: LOG") - model.h2o.poisson.log <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="poisson", link="log",alpha=0.5, lambda=0, n_folds=0) + model.h2o.poisson.log <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="poisson", link="log",alpha=0.5, lambda=0, nfolds=0) model.R.poisson.log <- glm(formula=R.formula, data=R.data[,2:9], family=poisson(link=log), na.action=na.omit) print("Compare model deviances for link function log") @@ -36,7 +36,7 @@ test.linkFunctions <- function(conn) { } print("Create models with link: IDENTITY") - model.h2o.poisson.identity <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="poisson", link="identity",alpha=0.5, lambda=0, n_folds=0) + model.h2o.poisson.identity <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="poisson", link="identity",alpha=0.5, lambda=0, nfolds=0) model.R.poisson.identity <- glm(formula=R.formula, data=R.data[,2:9], family=poisson(link=identity), na.action=na.omit) print("Compare model deviances for link function identity") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_balanced.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_balanced.R index 29b7051ee8db..bc5ffeb00656 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_balanced.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_balanced.R @@ -16,7 +16,7 @@ test <- function(conn) { data.b.hex <- h2o.uploadFile(conn, locate("smalldata/synthetic_perfect_separation/balanced.csv"), key="data.b.hex") print("Fit model on dataset.") - model.balanced <- h2o.glm(x=c("x1", "x2"), y="y", data.b.hex, family="binomial", lambda_search=TRUE, use_all_factor_levels=TRUE, alpha=0.5, n_folds=0, lambda=0) + model.balanced <- h2o.glm(x=c("x1", "x2"), y="y", data.b.hex, family="binomial", lambda_search=TRUE, use_all_factor_levels=TRUE, alpha=0.5, nfolds=0, lambda=0) print("Extract models' coefficients and assert reasonable values (ie. no greater than 50)") print("Balanced dataset") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_unbalanced.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_unbalanced.R index ed31667ec633..0ed95b162f61 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_unbalanced.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_perfectSeparation_unbalanced.R @@ -16,7 +16,7 @@ test <- function(conn) { data.u.hex <- h2o.uploadFile(conn, locate("smalldata/synthetic_perfect_separation/unbalanced.csv"), key="data.u.hex") print("Fit model on dataset.") - model.unbalanced <- h2o.glm(x=c("x1", "x2"), y="y", data.u.hex, family="binomial", lambda_search=TRUE, use_all_factor_levels=TRUE, alpha=0.5, n_folds=0, lambda=0) + model.unbalanced <- h2o.glm(x=c("x1", "x2"), y="y", data.u.hex, family="binomial", lambda_search=TRUE, use_all_factor_levels=TRUE, alpha=0.5, nfolds=0, lambda=0) print("Extract models' coefficients and assert reasonable values (ie. no greater than 50)") print("Unbalanced dataset") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_shuffling_large.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_shuffling_large.R index 4beb5b99677d..e2e00492338e 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_shuffling_large.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_shuffling_large.R @@ -17,13 +17,13 @@ test <- function(conn) { arcene.train.full_shuffled = h2o.assign(arcene.train.full[sample(nrow(arcene.train.full),replace=F),],"arcene.train.full_shuffled") print("Create model on original Arcene dataset.") - h2o.model <- h2o.glm(x=c(1:1000), y=1001, training_frame=arcene.train.full, family="binomial", lambda_search=TRUE, alpha=0.5, n_folds=0, use_all_factor_levels=TRUE) + h2o.model <- h2o.glm(x=c(1:1000), y=1001, training_frame=arcene.train.full, family="binomial", lambda_search=TRUE, alpha=0.5, nfolds=0, use_all_factor_levels=TRUE) print("Create second model on original Arcene dataset.") - h2o.model2 <- h2o.glm(x=c(1:1000), y=1001, training_frame=arcene.train.full, family="binomial", lambda_search=TRUE, alpha=0.5, n_folds=0, use_all_factor_levels=TRUE) + h2o.model2 <- h2o.glm(x=c(1:1000), y=1001, training_frame=arcene.train.full, family="binomial", lambda_search=TRUE, alpha=0.5, nfolds=0, use_all_factor_levels=TRUE) print("Create model on shuffled Arcene dataset.") - h2o.model.s <- h2o.glm(x=c(1:1000), y=1001, training_frame=arcene.train.full_shuffled, family="binomial", lambda_search=TRUE, alpha=0.5, n_folds=0, use_all_factor_levels=TRUE) + h2o.model.s <- h2o.glm(x=c(1:1000), y=1001, training_frame=arcene.train.full_shuffled, family="binomial", lambda_search=TRUE, alpha=0.5, nfolds=0, use_all_factor_levels=TRUE) print("Assert that number of predictors remaining and their respective coefficients are equal.") print("Comparing 2 models from original dataset") diff --git a/h2o-r/tests/testdir_algos/glm/runit_GLM_wide_dataset_large.R b/h2o-r/tests/testdir_algos/glm/runit_GLM_wide_dataset_large.R index a2990ab35f5e..5a7314b129d1 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_GLM_wide_dataset_large.R +++ b/h2o-r/tests/testdir_algos/glm/runit_GLM_wide_dataset_large.R @@ -25,7 +25,7 @@ test <- function(conn) { arcene.valid.full = h2o.assign(data=h2o.cbind(arcene.valid,arcene.valid.label),key="arcene.valid.full") print("Run model on 3250 columns of Arcene with strong rules off.") - time.noSR.3250 <- system.time(model.noSR.3250 <- h2o.glm(x=c(1:3250), y="arcene.train.label", training_frame=arcene.train.full, family="binomial", lambda_search=FALSE, alpha=1, n_folds=0, use_all_factor_levels=TRUE)) + time.noSR.3250 <- system.time(model.noSR.3250 <- h2o.glm(x=c(1:3250), y="arcene.train.label", training_frame=arcene.train.full, family="binomial", lambda_search=FALSE, alpha=1, nfolds=0, use_all_factor_levels=TRUE)) print("Test model on validation set.") predict.noSR.3250 <- predict(model.noSR.3250, arcene.valid.full) diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLMGrid_lambda_search.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLMGrid_lambda_search.R index 6ef660e6ebbb..7a91db70c077 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLMGrid_lambda_search.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLMGrid_lambda_search.R @@ -9,7 +9,7 @@ test.GLMGrid.lambda.search <- function(conn) { Log.info("H2O GLM (binomial) with parameters: alpha = c(0.25, 0.5), nlambda = 20, lambda_search = TRUE, nfolds: 2\n") # missing alpha=c(0.25, 0.5) - prostate.bestlambda = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = TRUE, n_folds = 2) + prostate.bestlambda = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = TRUE, nfolds = 2) model_idx = ifelse(runif(1) <= 0.5, 1, 2) model.bestlambda = prostate.bestlambda@model[[model_idx]] params.bestlambda = model.bestlambda@model$params @@ -35,7 +35,7 @@ test.GLMGrid.lambda.search <- function(conn) { expect_equal(best_model@model, model.bestlambda@model) Log.info("H2O GLM (binomial) with parameters: alpha = c(0.25, 0.5), nlambda = 20, lambda_search = TRUE, nfolds: 2\n") - prostate.search = h2o.glm(x = 3:9, y = 2, data = prostate.hex, family = "binomial", alpha = c(0.25, 0.5), nlambdas = 5, lambda_search = TRUE, n_folds = 2) + prostate.search = h2o.glm(x = 3:9, y = 2, data = prostate.hex, family = "binomial", alpha = c(0.25, 0.5), nlambdas = 5, lambda_search = TRUE, nfolds = 2) model.search = prostate.search@model[[model_idx]] models.best = model.search@models[[model.search@best_model]] params.best = models.best@model$params diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_benign.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_benign.R index 295345ce1bd9..6af2dc2344b9 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_benign.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_benign.R @@ -10,7 +10,7 @@ glm2Benign <- function(conn) { X <- X[ X != Y ] Log.info("Build the model") - mFV <- h2o.glm(y = Y, x = colnames(bhexFV)[X], training_frame = bhexFV, family = "binomial", n_folds = 5, alpha = 0, lambda = 1e-5) + mFV <- h2o.glm(y = Y, x = colnames(bhexFV)[X], training_frame = bhexFV, family = "binomial", nfolds = 5, alpha = 0, lambda = 1e-5) Log.info("Check that the columns used in the model are the ones we passed in.") diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype.R index 2f1c59b65ff4..359e65f65948 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype.R @@ -19,21 +19,21 @@ test.GLM.covtype <- function(conn) { # L2: alpha = 0, lambda = 0 start = Sys.time() - covtype.h2o1 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", n_folds = 2, alpha = 0, lambda = 0) + covtype.h2o1 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", nfolds = 2, alpha = 0, lambda = 0) end = Sys.time() Log.info(cat("GLM (L2) on", covtype.hex@key, "took", as.numeric(end-start), "seconds\n")) print(covtype.h2o1) # Elastic: alpha = 0.5, lambda = 1e-4 start = Sys.time() - covtype.h2o2 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", n_folds = 2, alpha = 0.5, lambda = 1e-4) + covtype.h2o2 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", nfolds = 2, alpha = 0.5, lambda = 1e-4) end = Sys.time() Log.info(cat("GLM (Elastic) on", covtype.hex@key, "took", as.numeric(end-start), "seconds\n")) print(covtype.h2o2) # L1: alpha = 1, lambda = 1e-4 start = Sys.time() - covtype.h2o3 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", n_folds = 2, alpha = 1, lambda = 1e-4) + covtype.h2o3 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", nfolds = 2, alpha = 1, lambda = 1e-4) end = Sys.time() Log.info(cat("GLM (L1) on", covtype.hex@key, "took", as.numeric(end-start), "seconds\n")) print(covtype.h2o3) diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype_getModel.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype_getModel.R index f25134c98814..c065f39f2fa2 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype_getModel.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_covtype_getModel.R @@ -19,7 +19,7 @@ test.GLM.covtype <- function(conn) { # L2: alpha = 0, lambda = 0 start = Sys.time() - covtype.h2o1 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", n_folds = 2, alpha = 0, lambda = 0) + covtype.h2o1 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", nfolds = 2, alpha = 0, lambda = 0) end = Sys.time() Log.info(cat("GLM (L2) on", covtype.hex@key, "took", as.numeric(end-start), "seconds\n")) print(covtype.h2o1) @@ -28,7 +28,7 @@ test.GLM.covtype <- function(conn) { # Elastic: alpha = 0.5, lambda = 1e-4 start = Sys.time() - covtype.h2o2 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", n_folds = 2, alpha = 0.5, lambda = 1e-4) + covtype.h2o2 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", nfolds = 2, alpha = 0.5, lambda = 1e-4) end = Sys.time() Log.info(cat("GLM (Elastic) on", covtype.hex@key, "took", as.numeric(end-start), "seconds\n")) print(covtype.h2o2) @@ -37,7 +37,7 @@ test.GLM.covtype <- function(conn) { # L1: alpha = 1, lambda = 1e-4 start = Sys.time() - covtype.h2o3 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", n_folds = 2, alpha = 1, lambda = 1e-4) + covtype.h2o3 = h2o.glm(y = myY, x = myX, training_frame = covtype.hex, family = "binomial", nfolds = 2, alpha = 1, lambda = 1e-4) end = Sys.time() Log.info(cat("GLM (L1) on", covtype.hex@key, "took", as.numeric(end-start), "seconds\n")) print(covtype.h2o3) diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_lambda_search_large.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_lambda_search_large.R index 063b83a2bed4..aeb4c1dc6fd7 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_lambda_search_large.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_lambda_search_large.R @@ -9,13 +9,13 @@ test.GLM.lambda.search <- function(conn) { # GLM without lambda search, lambda is single user-provided value Log.info("H2O GLM (binomial) with parameters: lambda_search = TRUE, nfolds: 2\n") - prostate.nosearch = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = FALSE, n_folds = 2) + prostate.nosearch = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = FALSE, nfolds = 2) params.nosearch = prostate.nosearch@model$params expect_error(h2o.getGLMLambdaModel(prostate.nosearch, 0.5)) # GLM with lambda search, return only model corresponding to best lambda as determined by H2O Log.info("H2O GLM (binomial) with parameters: lambda_search: TRUE, nfolds: 2\n") - prostate.bestlambda = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = TRUE, n_folds = 2) + prostate.bestlambda = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = TRUE, nfolds = 2) params.bestlambda = prostate.bestlambda@model$params random_lambda = sample(params.bestlambda$lambda_all, 1) @@ -29,7 +29,7 @@ test.GLM.lambda.search <- function(conn) { # GLM with lambda search, return models corresponding to all lambda searched over Log.info("H2O GLM (binomial) with parameters: lambda_search: TRUE, nfolds: 2\n") - prostate.search = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = TRUE, n_folds = 2) + prostate.search = h2o.glm(x = 3:9, y = 2, training_frame = prostate.hex, family = "binomial", nlambdas = 5, lambda_search = TRUE, nfolds = 2) models.best = prostate.search@models[[prostate.search@best_model]] models.bestlambda = models.best@model$params$lambda_best expect_equal(models.best@model$lambda, models.bestlambda) diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_link_functions_tweedie.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_link_functions_tweedie.R index 896c9ea63c40..97ba348b0c22 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_link_functions_tweedie.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_link_functions_tweedie.R @@ -21,7 +21,7 @@ test.linkFunctions <- function(conn) { R.formula = (R.data[,"CAPSULE"]~.) print("Create models with canonical link: TWEEDIE") - model.h2o.tweedie.tweedie <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="tweedie", link="tweedie",alpha=0.5, lambda=0, n_folds=0) + model.h2o.tweedie.tweedie <- h2o.glm(x=myX, y=myY, training_frame=h2o.data, family="tweedie", link="tweedie",alpha=0.5, lambda=0, nfolds=0) model.R.tweedie.tweedie <- glm(formula=R.formula, data=R.data[,4:10], family=tweedie, na.action=na.omit) print("Compare model deviances for link function tweedie") diff --git a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_prostate.R b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_prostate.R index 87874ee81476..ca9f10bfc689 100644 --- a/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_prostate.R +++ b/h2o-r/tests/testdir_algos/glm/runit_NOFEATURE_GLM_prostate.R @@ -17,7 +17,7 @@ test.GLM.prostate <- function(conn) { # myX = paste(myX, collapse=",") Log.info(cat("B)H2O GLM (binomial) with parameters:\nX:", myX, "\nY:", myY, "\n")) - prostate.glm.h2o = h2o.glm(y = myY, x = myX, training_frame = prostate.hex, family = "binomial", n_folds = 10, alpha = 0.5) + prostate.glm.h2o = h2o.glm(y = myY, x = myX, training_frame = prostate.hex, family = "binomial", nfolds = 10, alpha = 0.5) print(prostate.glm.h2o) # prostate.glm = glm.fit(y = prostate.data[,myY], x = prostate.data[,myX], family = binomial) diff --git a/h2o-r/tests/testdir_demos/runit_demo_tableau.R b/h2o-r/tests/testdir_demos/runit_demo_tableau.R index 9c8c91cda625..2d8f68063631 100644 --- a/h2o-r/tests/testdir_demos/runit_demo_tableau.R +++ b/h2o-r/tests/testdir_demos/runit_demo_tableau.R @@ -32,7 +32,7 @@ test.tableau <- function(conn) { .arg2 <- 'Origin,Dest,UniqueCarrier' xvars <- unlist( strsplit( .arg2, split = ',' , fixed = TRUE ) ) - data.glm <- h2o.glm(x = xvars , y = 'Cancelled', training_frame = data.hex, family = 'binomial', n_folds = 0, standardize=TRUE) + data.glm <- h2o.glm(x = xvars , y = 'Cancelled', training_frame = data.hex, family = 'binomial', nfolds = 0, standardize=TRUE) glmModelTemp <- eval(parse(text = 'data.glm' )) originFactors <- levels(data.hex$Origin) diff --git a/h2o-r/tests/testdir_demos/runit_demo_tk_cm_roc.R b/h2o-r/tests/testdir_demos/runit_demo_tk_cm_roc.R index bcd9f026f2df..8445076084bd 100644 --- a/h2o-r/tests/testdir_demos/runit_demo_tk_cm_roc.R +++ b/h2o-r/tests/testdir_demos/runit_demo_tk_cm_roc.R @@ -94,6 +94,6 @@ if (FALSE) { head(df) myX <- c("Origin", "Dest", "Distance", "UniqueCarrier", "Month", "DayofMonth", "DayOfWeek") myY <- "IsDepDelayed" - air.glm <- h2o.glm(x = myX, y = myY, training_frame = df, family = "binomial", n_folds = 10, alpha = 0.25, lambda = 0.001) + air.glm <- h2o.glm(x = myX, y = myY, training_frame = df, family = "binomial", nfolds = 10, alpha = 0.25, lambda = 0.001) air.glm@model$confusion } diff --git a/h2o-r/tests/testdir_docexamples/runit_NOPASS_Rdoc_glm.R b/h2o-r/tests/testdir_docexamples/runit_NOPASS_Rdoc_glm.R index 710d171a7fcc..94c04bf666db 100644 --- a/h2o-r/tests/testdir_docexamples/runit_NOPASS_Rdoc_glm.R +++ b/h2o-r/tests/testdir_docexamples/runit_NOPASS_Rdoc_glm.R @@ -6,9 +6,9 @@ test.rdocglm.golden <- function(H2Oserver) { prostate.hex <- h2o.importURL(H2Oserver, path = locate("smalldata/logreg/prostate.csv"), key = "prostate.hex") -h2o.glm(y = "CAPSULE", x = c("AGE","RACE","PSA","DCAPS"), training_frame = prostate.hex, family = "binomial", n_folds = 10, alpha = 0.5) +h2o.glm(y = "CAPSULE", x = c("AGE","RACE","PSA","DCAPS"), training_frame = prostate.hex, family = "binomial", nfolds = 10, alpha = 0.5) myX <- setdiff(colnames(prostate.hex), c("ID", "DPROS", "DCAPS", "VOL")) -h2o.glm(y = "VOL", x = myX, training_frame = prostate.hex, family = "gaussian", n_folds = 5, alpha = 0.1) +h2o.glm(y = "VOL", x = myX, training_frame = prostate.hex, family = "gaussian", nfolds = 5, alpha = 0.1) airlines.hex <- h2o.importURL(H2Oserver, path = locate("smalldata/airlines/AirlinesTrain.csv.zip")) h2o.glm(x = c('Distance', 'Origin', 'Dest', 'UniqueCarrier'), y = 'IsDepDelayed', family = 'binomial', training_frame = airlines.hex) diff --git a/h2o-r/tests/testdir_golden/runitP_NOPASS_glm2_5_golden.R b/h2o-r/tests/testdir_golden/runitP_NOPASS_glm2_5_golden.R index d6e80981c0ae..9d1c3bc5c9e5 100644 --- a/h2o-r/tests/testdir_golden/runitP_NOPASS_glm2_5_golden.R +++ b/h2o-r/tests/testdir_golden/runitP_NOPASS_glm2_5_golden.R @@ -9,7 +9,7 @@ test.glm2ProstateAUC.golden <- function(H2Oserver) { prostateR<- read.csv(locate("smalldata/logreg/prostate.csv"), header=T) Log.info("Run matching models in R and H2O") - fitH2O<- h2o.glm(y="CAPSULE", x=c("AGE", "RACE", "DPROS", "DCAPS", "PSA", "VOL", "GLEASON"), training_frame=prostateH2O, family="binomial", lambda=0, alpha=0, n_folds=0, standardize=F) + fitH2O<- h2o.glm(y="CAPSULE", x=c("AGE", "RACE", "DPROS", "DCAPS", "PSA", "VOL", "GLEASON"), training_frame=prostateH2O, family="binomial", lambda=0, alpha=0, nfolds=0, standardize=F) fitR<- glm(CAPSULE ~ AGE + RACE + DPROS + DCAPS + PSA + VOL + GLEASON, family=binomial, data=prostateR) prostateR$predsR<- predict.glm(fitR, newdata=NULL, type="response") preds2R<- prediction(prostateR$predsR, labels=prostateR$CAPSULE) diff --git a/h2o-r/tests/testdir_golden/runit_NOPASS_glm2_10_golden.R b/h2o-r/tests/testdir_golden/runit_NOPASS_glm2_10_golden.R index 5ba6c9b731bb..e3f49b6cb802 100644 --- a/h2o-r/tests/testdir_golden/runit_NOPASS_glm2_10_golden.R +++ b/h2o-r/tests/testdir_golden/runit_NOPASS_glm2_10_golden.R @@ -24,7 +24,7 @@ ridgeGLMNet <- function (X,y,L){ #H2O RIDGE ridgeH2O <- function (X,y,L){ - fitH2O=h2o.glm(X, y, training_frame=handmadeH2O, n_folds=0, alpha=0, lambda=L, family="gaussian", standardize=T) + fitH2O=h2o.glm(X, y, training_frame=handmadeH2O, nfolds=0, alpha=0, lambda=L, family="gaussian", standardize=T) betah <- fitH2O@model$coefficients_table$'Norm Coefficients'[hX] betah <- c(betah, fitH2O@model$coefficients_table$'Norm Coefficients'[length(fitH2O@model$coefficients_table$'Norm Coefficients')]) diff --git a/h2o-r/tests/testdir_golden/runit_glm2_11_golden.R b/h2o-r/tests/testdir_golden/runit_glm2_11_golden.R index e523b127a7eb..7eaa33e1a796 100644 --- a/h2o-r/tests/testdir_golden/runit_glm2_11_golden.R +++ b/h2o-r/tests/testdir_golden/runit_glm2_11_golden.R @@ -18,7 +18,7 @@ function(conn) { lambda = 1e-5 #H2O GLM model - hh=h2o.glm(x=myX,y=myY,training_frame=mfrmr,family="gaussian",n_folds=0, alpha = alpha, lambda = lambda) + hh=h2o.glm(x=myX,y=myY,training_frame=mfrmr,family="gaussian",nfolds=0, alpha = alpha, lambda = lambda) res_dev = hh@model$residual_deviance obs = nrow(mfrmr) diff --git a/h2o-r/tests/testdir_golden/runit_glm2_1_golden.R b/h2o-r/tests/testdir_golden/runit_glm2_1_golden.R index 81948a094246..ba047af6c89e 100644 --- a/h2o-r/tests/testdir_golden/runit_glm2_1_golden.R +++ b/h2o-r/tests/testdir_golden/runit_glm2_1_golden.R @@ -10,7 +10,7 @@ cuseR <- read.csv(locate("smalldata/logreg/cuseexpanded.csv"), header=T) Log.info("Test H2O treatment of FACTORS AS PREDICTORS") Log.info("Run matching models in R and H2O") -fitH2O <- h2o.glm(y="UsingBinom", x=c("Age", "Ed", "Wantsmore"), training_frame=cuseH2O, family="binomial", lambda=0, alpha=0, n_folds=0) +fitH2O <- h2o.glm(y="UsingBinom", x=c("Age", "Ed", "Wantsmore"), training_frame=cuseH2O, family="binomial", lambda=0, alpha=0, nfolds=0) fitR <- glm(UsingBinom ~ AgeA + AgeC + AgeD + LowEd + MoreYes, family=binomial, data=cuseR) diff --git a/h2o-r/tests/testdir_golden/runit_glm2_2_golden.R b/h2o-r/tests/testdir_golden/runit_glm2_2_golden.R index df3cf3c96828..83430cc13604 100644 --- a/h2o-r/tests/testdir_golden/runit_glm2_2_golden.R +++ b/h2o-r/tests/testdir_golden/runit_glm2_2_golden.R @@ -9,7 +9,7 @@ test.glm2Prostate.golden <- function(H2Oserver) { prostateR<- read.csv(locate("smalldata/logreg/prostate.csv"), header=T) Log.info("Run matching models in R and H2O") - fitH2O<- h2o.glm(y="CAPSULE", x=c("AGE", "RACE", "DPROS", "DCAPS", "PSA", "VOL", "GLEASON"), training_frame=prostateH2O, family="binomial", lambda=0, alpha=0, n_folds=0, standardize=F) + fitH2O<- h2o.glm(y="CAPSULE", x=c("AGE", "RACE", "DPROS", "DCAPS", "PSA", "VOL", "GLEASON"), training_frame=prostateH2O, family="binomial", lambda=0, alpha=0, nfolds=0, standardize=F) fitR<- glm(CAPSULE ~ AGE + RACE + DPROS + DCAPS + PSA + VOL + GLEASON, family=binomial, data=prostateR) diff --git a/h2o-r/tests/testdir_golden/runit_glm2_3_golden.R b/h2o-r/tests/testdir_golden/runit_glm2_3_golden.R index 908f30d7710f..8ada3c3b466a 100644 --- a/h2o-r/tests/testdir_golden/runit_glm2_3_golden.R +++ b/h2o-r/tests/testdir_golden/runit_glm2_3_golden.R @@ -10,7 +10,7 @@ cuseR<- read.csv(locate("smalldata/logreg/cuseexpanded.csv"), header=T) Log.info("Test H2O Poisson not regularized") Log.info("Run matching models in R and H2O") -fitH2O<- h2o.glm(y="Using", x=c("Age", "Ed", "Wantsmore"), training_frame=cuseH2O, family="poisson", lambda=0, alpha=0, n_folds=0) +fitH2O<- h2o.glm(y="Using", x=c("Age", "Ed", "Wantsmore"), training_frame=cuseH2O, family="poisson", lambda=0, alpha=0, nfolds=0) fitR<- glm(Using ~ AgeA + AgeC + AgeD + LowEd + MoreYes, family=poisson, data=cuseR) diff --git a/h2o-r/tests/testdir_golden/runit_glm2_4_golden.R b/h2o-r/tests/testdir_golden/runit_glm2_4_golden.R index 23fa73aa551d..f124912e7a33 100644 --- a/h2o-r/tests/testdir_golden/runit_glm2_4_golden.R +++ b/h2o-r/tests/testdir_golden/runit_glm2_4_golden.R @@ -40,7 +40,7 @@ RT1<- ridgeLinear(x, y, L) #fit corresponding H2O model -fitH2O<- h2o.glm(x=c("V8", "V9", "V10", "V11", "V12"), y="V13", family="gaussian", n_folds=0, alpha=0, lambda=0.01, training_frame=hmH2O) +fitH2O<- h2o.glm(x=c("V8", "V9", "V10", "V11", "V12"), y="V13", family="gaussian", nfolds=0, alpha=0, lambda=0.01, training_frame=hmH2O) #test that R coefficients and basic descriptives are equal Rcoeffsglmnet<- sort(as.matrix(coefficients(fitRglmnet))) diff --git a/h2o-r/tests/testdir_golden/runit_glm2_objectiveFun_golden.R b/h2o-r/tests/testdir_golden/runit_glm2_objectiveFun_golden.R index d4f0c2fbd3dc..4bbf0a150e1a 100644 --- a/h2o-r/tests/testdir_golden/runit_glm2_objectiveFun_golden.R +++ b/h2o-r/tests/testdir_golden/runit_glm2_objectiveFun_golden.R @@ -18,7 +18,7 @@ function(conn) { lambda <- 1e-5 # H2O GLM model - hh <- h2o.glm(x=myX,y=myY,training_frame=mfrmr,family="gaussian",n_folds=0, alpha = alpha, lambda = lambda) + hh <- h2o.glm(x=myX,y=myY,training_frame=mfrmr,family="gaussian",nfolds=0, alpha = alpha, lambda = lambda) res_dev <- hh@model$residual_deviance obs <- nrow(mfrmr) # lambda <- hh@model$params$lambda_best diff --git a/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1775_save_load.R b/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1775_save_load.R index 06efadb2c028..0724130f3028 100644 --- a/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1775_save_load.R +++ b/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1775_save_load.R @@ -23,7 +23,7 @@ test.hex_1775 <- function(conn) { # Build GLM, RandomForest, GBM, Naive Bayes, and Deep Learning models Log.info("Build GLM model") - prostate.glm = h2o.glm(y = "CAPSULE", x = c("AGE","RACE","PSA","DCAPS"), training_frame = prostate.hex, family = "binomial", n_folds = 0, alpha = 0.5) + prostate.glm = h2o.glm(y = "CAPSULE", x = c("AGE","RACE","PSA","DCAPS"), training_frame = prostate.hex, family = "binomial", nfolds = 0, alpha = 0.5) Log.info("Build GBM model") prostate.gbm = h2o.gbm(y = 2, x = 3:9, training_frame = prostate.hex, nfolds = 5, loss = "multinomial") Log.info("Build Speedy Random Forest Model") diff --git a/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1799_glm_nfold_parameters.R b/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1799_glm_nfold_parameters.R index 1ee2ee18fbd8..baf757d07bdd 100644 --- a/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1799_glm_nfold_parameters.R +++ b/h2o-r/tests/testdir_jira/runit_NOFEATURE_hex_1799_glm_nfold_parameters.R @@ -16,7 +16,7 @@ function(conn) { path <- locate("smalldata/logreg/prostate.csv") prostate.hex <- h2o.importFile(conn, path, key="prostate.hex") - main_model <- h2o.glm(x = 3:8, y = 2, training_frame = prostate.hex, n_folds = 2, standardize = FALSE, family = "binomial") + main_model <- h2o.glm(x = 3:8, y = 2, training_frame = prostate.hex, nfolds = 2, standardize = FALSE, family = "binomial") print(main_model@key) diff --git a/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_837_glm_assertion_large.R b/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_837_glm_assertion_large.R index 629138abe2f1..28616d759d19 100644 --- a/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_837_glm_assertion_large.R +++ b/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_837_glm_assertion_large.R @@ -26,12 +26,12 @@ test <- function(conn) { print("Creating model without CV") system.time(h2o.glm.model <- h2o.glm(x=myX, y=myY, training_frame=adult.train, destination_key="h2o.glm.adult", family="binomial", - alpha=1, lambda_search=T, n_folds=0, use_all_factor_levels=TRUE)) + alpha=1, lambda_search=T, nfolds=0, use_all_factor_levels=TRUE)) h2o.glm.model print("Creating model with CV") system.time(h2o.glm.CV <- h2o.glm(x=myX, y=myY, training_frame=adult.train, destination_key="h2o.glm.CV.adult", family="binomial", - alpha=1, lambda_search=T, n_folds=5, use_all_factor_levels=TRUE)) # This line is failing + alpha=1, lambda_search=T, nfolds=5, use_all_factor_levels=TRUE)) # This line is failing h2o.glm.CV testEnd() diff --git a/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_874_glm_cv_nonzero_reporting_large.R b/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_874_glm_cv_nonzero_reporting_large.R index bfec45e19e6c..e108598d3936 100644 --- a/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_874_glm_cv_nonzero_reporting_large.R +++ b/h2o-r/tests/testdir_jira/runit_NOFEATURE_pub_874_glm_cv_nonzero_reporting_large.R @@ -18,7 +18,7 @@ test <- function(conn) { myY <- "label" print("Creating model with CV") h2o.glm.CV <- h2o.glm(x=myX, y=myY, training_frame=mushroom.train, destination_key="h2o.glm.CV.mushroom", family="binomial", - alpha=1, lambda_search=T, n_folds=3, use_all_factor_levels=TRUE) + alpha=1, lambda_search=T, nfolds=3, use_all_factor_levels=TRUE) print(h2o.glm.CV) #Confirm reported values accurate and match browser print("Reading in Abalone data for gaussian glm.") @@ -27,7 +27,7 @@ test <- function(conn) { myY <- "C9" print("Creating model with CV") h2o.glm.CV <- h2o.glm(x=myX, y=myY, training_frame=abalone.train, destination_key="h2o.glm.CV.abalone", family="gaussian", - alpha=1, lambda_search=T, n_folds=3, use_all_factor_levels=TRUE) + alpha=1, lambda_search=T, nfolds=3, use_all_factor_levels=TRUE) print(h2o.glm.CV) #Confirm reported values accurate and match browser testEnd() diff --git a/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1896_glm_intercepts.R b/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1896_glm_intercepts.R index c644765c8fda..762d725bd90f 100644 --- a/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1896_glm_intercepts.R +++ b/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1896_glm_intercepts.R @@ -44,14 +44,14 @@ test.GLM.zero_intercept <- function(conn) { Log.info("Build logistic model in H2O with intercept...") prostate.glm.h2o1 = h2o.glm(y = myY, x = myX, training_frame = prostate.hex, lambda = var_lambda1, - family = var_family, n_folds = var_folds, alpha = var_alpha, intercept = TRUE) + family = var_family, nfolds = var_folds, alpha = var_alpha, intercept = TRUE) Log.info("Build logistic model in H2O without intercept...") ## standardization must be set to false since there are no intercepts, we cannnot regularize prostate.glm.h2o2 = h2o.glm(y = myY, x = myX, training_frame = prostate.hex, lambda = var_lambda2, standardize = F, - family = var_family, n_folds = var_folds, alpha = var_alpha, intercept = FALSE) + family = var_family, nfolds = var_folds, alpha = var_alpha, intercept = FALSE) Log.info("Build logistic model in H2O w/o intercept w/ rebalanced data...") prostate.glm.h2o3 = h2o.glm(y = myY, x = myX, training_frame = prostate.rebalanced, lambda = var_lambda2, standardize = F, - family = var_family, n_folds = var_folds, alpha = var_alpha, intercept = FALSE) + family = var_family, nfolds = var_folds, alpha = var_alpha, intercept = FALSE) check_coeff(prostate.glm.h2o2@model$coefficients, prostate.glm.h2o3@model$coefficients, 1e-10) Log.info("Rebalanced data ran with same results.") diff --git a/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1908_save_load_all.R b/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1908_save_load_all.R index 4013bb6715e7..2901a6449bdb 100644 --- a/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1908_save_load_all.R +++ b/h2o-r/tests/testdir_jira/runit_NOPASS_hex_1908_save_load_all.R @@ -25,11 +25,11 @@ test.hex_1908 <- function(conn) { # Build GLM and GBM each with cross validation models Log.info("Build GLM model") - airlines.glm = h2o.glm(y = myY, x = myX, training_frame = airlines.hex, family = "binomial", n_folds = 0, alpha = 0.5) + airlines.glm = h2o.glm(y = myY, x = myX, training_frame = airlines.hex, family = "binomial", nfolds = 0, alpha = 0.5) Log.info("Build GLM model with nfold = 5") - airlines_xval.glm = h2o.glm(y = myY, x = myX, training_frame = airlines.hex, family = "binomial", n_folds = 5, alpha = 0.5) + airlines_xval.glm = h2o.glm(y = myY, x = myX, training_frame = airlines.hex, family = "binomial", nfolds = 5, alpha = 0.5) Log.info("Build GBM model with nfold = 3") - airlines_xval.gbm = h2o.gbm(y = myY, x = myX, training_frame = airlines.hex, n_folds = 3, loss = "multinomial") + airlines_xval.gbm = h2o.gbm(y = myY, x = myX, training_frame = airlines.hex, nfolds = 3, loss = "multinomial") # Predict on models and save results in R Log.info("Scoring on models and saving predictions to R") diff --git a/h2o-r/tests/testdir_jira/runit_NOPASS_pub_960_glm_aic.R b/h2o-r/tests/testdir_jira/runit_NOPASS_pub_960_glm_aic.R index 34c943de515b..36f07ed35197 100644 --- a/h2o-r/tests/testdir_jira/runit_NOPASS_pub_960_glm_aic.R +++ b/h2o-r/tests/testdir_jira/runit_NOPASS_pub_960_glm_aic.R @@ -21,13 +21,13 @@ test <- function(conn) { print("Testing for family: GAMMA") model.h2o.Gamma.inverse <- h2o.glm(x=myX, y=myY, training_frame=prostate.data, family="gamma", link="inverse", alpha=1, lambda_search=T, # variable_importance=TRUE, - use_all_factor_levels=TRUE, n_folds=0) + use_all_factor_levels=TRUE, nfolds=0) print(model.h2o.Gamma.inverse) #AIC is NaN print("Testing for family: TWEEDIE") model.h2o.tweedie <- h2o.glm(x=myX, y=myY, training_frame=prostate.data, family="tweedie", link="tweedie", alpha=1, lambda_search=T, # variable_importance=TRUE, - use_all_factor_levels=TRUE, n_folds=0) + use_all_factor_levels=TRUE, nfolds=0) print(model.h2o.tweedie) #AIC is NaN testEnd() diff --git a/h2o-r/tests/testdir_jira/runit_NOPASS_pub_965_binomial_log_pred.R b/h2o-r/tests/testdir_jira/runit_NOPASS_pub_965_binomial_log_pred.R index 4d900d4b2903..90fee6c73727 100644 --- a/h2o-r/tests/testdir_jira/runit_NOPASS_pub_965_binomial_log_pred.R +++ b/h2o-r/tests/testdir_jira/runit_NOPASS_pub_965_binomial_log_pred.R @@ -24,7 +24,7 @@ test.linkFunctions <- function(conn) { myX = c("AGE","RACE","DCAPS","PSA","VOL","DPROS","GLEASON") print("Create model with link: LOG") - model.h2o.binomial.log <- h2o.glm(x=myX, y=myY, training_frame=prostate.train, family="binomial", link="log",alpha=0.5, lambda=0, n_folds=0) + model.h2o.binomial.log <- h2o.glm(x=myX, y=myY, training_frame=prostate.train, family="binomial", link="log",alpha=0.5, lambda=0, nfolds=0) print("Predict") prediction.h2o.binomial.log <- predict(model.h2o.binomial.log, prostate.test) diff --git a/h2o-r/tests/testdir_jira/runit_hex_1750_strongRules_mem.R b/h2o-r/tests/testdir_jira/runit_hex_1750_strongRules_mem.R index 7876a28fee67..e731997087a4 100644 --- a/h2o-r/tests/testdir_jira/runit_hex_1750_strongRules_mem.R +++ b/h2o-r/tests/testdir_jira/runit_hex_1750_strongRules_mem.R @@ -27,7 +27,7 @@ test <- function(conn) { # assertError(H2OModel.noSR.fail <- h2o.glm(x=c(1:7000), y="arcene.train.label", data=arcene.train.full, family="binomial", lambda_search=FALSE,alpha=1, nfolds=0, use_all_factor_levels=1)) print("Model successfully created for arcene data with 7000 columns with strong rules on.") - H2OModel.SR.pass <- h2o.glm(x=c(1:7000), y="arcene.train.label", training_frame=arcene.train.full, family="binomial", lambda_search=T,alpha=1, n_folds=0, use_all_factor_levels=TRUE) + H2OModel.SR.pass <- h2o.glm(x=c(1:7000), y="arcene.train.label", training_frame=arcene.train.full, family="binomial", lambda_search=T,alpha=1, nfolds=0, use_all_factor_levels=TRUE) testEnd() } diff --git a/h2o-r/tests/testdir_jira/runit_pub_831_synthetic_strongRules.R b/h2o-r/tests/testdir_jira/runit_pub_831_synthetic_strongRules.R index 398671754d77..1b287ef6250f 100644 --- a/h2o-r/tests/testdir_jira/runit_pub_831_synthetic_strongRules.R +++ b/h2o-r/tests/testdir_jira/runit_pub_831_synthetic_strongRules.R @@ -32,25 +32,25 @@ test <- function(conn) { # Test that run time is within reasonable timeframe (ie. 30 seconds) # GLM aborted if exceeds time frame and test fails startTime <- proc.time() - prostate.def.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, n_folds=0) + prostate.def.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, nfolds=0) endTime <- proc.time() elapsedTime <- endTime - startTime stopifnot(elapsedTime < 60) startTime <- proc.time() - prostate.bin.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL","BIN"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, n_folds=0) + prostate.bin.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL","BIN"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, nfolds=0) endTime <- proc.time() elapsedTime <- endTime - startTime stopifnot(elapsedTime < 60) startTime <- proc.time() - prostate.float.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL","FLOAT"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, n_folds=0) + prostate.float.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL","FLOAT"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, nfolds=0) endTime <- proc.time() elapsedTime <- endTime - startTime stopifnot(elapsedTime < 60) startTime <- proc.time() - prostate.int.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL","INT"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, n_folds=0) + prostate.int.model <- h2o.glm(x=c("ID","CAPSULE","AGE","RACE","DPROS","DCAPS","PSA","VOL","INT"), y=c("GLEASON"), prostate.train, family="gaussian", lambda_search=FALSE, alpha=1, use_all_factor_levels=TRUE, nfolds=0) endTime <- proc.time() elapsedTime <- endTime - startTime stopifnot(elapsedTime < 60) diff --git a/h2o-r/tests/testdir_jira/runit_pub_838_h2o_perf_message.R b/h2o-r/tests/testdir_jira/runit_pub_838_h2o_perf_message.R index a3310886019d..fa50c0c697d1 100644 --- a/h2o-r/tests/testdir_jira/runit_pub_838_h2o_perf_message.R +++ b/h2o-r/tests/testdir_jira/runit_pub_838_h2o_perf_message.R @@ -23,7 +23,7 @@ test <- function(conn) { myY <- 2 print("Creating model") - system.time(h2o.glm.model <- h2o.glm(x=myX, y=myY, training_frame=prostate.train, destination_key="h2o.glm.prostate", family="binomial", alpha=1, lambda_search=F, n_folds=0, use_all_factor_levels=FALSE)) + system.time(h2o.glm.model <- h2o.glm(x=myX, y=myY, training_frame=prostate.train, destination_key="h2o.glm.prostate", family="binomial", alpha=1, lambda_search=F, nfolds=0, use_all_factor_levels=FALSE)) print("Predict on test data") prediction <- predict(h2o.glm.model, prostate.test) diff --git a/h2o-test-integ/tests/flow/100KRows2-5Cols.flow b/h2o-test-integ/tests/flow/100KRows2-5Cols.flow index e71e157ec50e..c73929a11408 100644 --- a/h2o-test-integ/tests/flow/100KRows2-5Cols.flow +++ b/h2o-test-integ/tests/flow/100KRows2-5Cols.flow @@ -35,11 +35,11 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-83f9f370-ddc7-449f-895c-878a0d4152af\",\"training_frame\":\"WU_100KRows2.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"ADMM\",\"max_iterations\":50,\"beta_eps\":0,\"standardize\":true,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-83f9f370-ddc7-449f-895c-878a0d4152af\",\"training_frame\":\"WU_100KRows2.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"ADMM\",\"max_iterations\":50,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"gaussian\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-e6b9862a-9163-4247-bca7-1bf4a670eac2\",\"training_frame\":\"WU_100KRows2.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-4219873518108646400,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-e6b9862a-9163-4247-bca7-1bf4a670eac2\",\"training_frame\":\"WU_100KRows2.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-4219873518108646400,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" }, { "type": "cs", diff --git a/h2o-test-integ/tests/flow/100KRows3KCols.flow b/h2o-test-integ/tests/flow/100KRows3KCols.flow index aa1360dcc7a0..7f421494689b 100644 --- a/h2o-test-integ/tests/flow/100KRows3KCols.flow +++ b/h2o-test-integ/tests/flow/100KRows3KCols.flow @@ -35,7 +35,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-f508fbfc-907c-4a39-b091-e64c5514c1ad\",\"training_frame\":\"WU_100KRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"L_BFGS\",\"max_iterations\":50,\"beta_eps\":0,\"standardize\":true,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-f508fbfc-907c-4a39-b091-e64c5514c1ad\",\"training_frame\":\"WU_100KRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"L_BFGS\",\"max_iterations\":50,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"gaussian\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", @@ -59,7 +59,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-567e1a58-990b-4031-be7f-ae817df5f297\",\"training_frame\":\"WU_100KRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-1371736214016650000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-567e1a58-990b-4031-be7f-ae817df5f297\",\"training_frame\":\"WU_100KRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-1371736214016650000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" }, { "type": "cs", diff --git a/h2o-test-integ/tests/flow/1MRows3KCols.flow b/h2o-test-integ/tests/flow/1MRows3KCols.flow index 74fd1f49924a..ae40ad277da2 100644 --- a/h2o-test-integ/tests/flow/1MRows3KCols.flow +++ b/h2o-test-integ/tests/flow/1MRows3KCols.flow @@ -35,7 +35,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-6f37cc1d-9128-42fc-a964-adce52f47f64\",\"training_frame\":\"WU_1MRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"L_BFGS\",\"max_iterations\":50,\"beta_eps\":0,\"standardize\":true,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-6f37cc1d-9128-42fc-a964-adce52f47f64\",\"training_frame\":\"WU_1MRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"L_BFGS\",\"max_iterations\":50,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"gaussian\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", @@ -59,7 +59,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-6eb86947-0156-41ce-918d-a45d1fc74d0a\",\"training_frame\":\"WU_1MRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":7706863078196985000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-6eb86947-0156-41ce-918d-a45d1fc74d0a\",\"training_frame\":\"WU_1MRows3KCols.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":7706863078196985000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" }, { "type": "cs", diff --git a/h2o-test-integ/tests/flow/BigCross.flow b/h2o-test-integ/tests/flow/BigCross.flow index b1882f3dd53c..7e9763ff7764 100644 --- a/h2o-test-integ/tests/flow/BigCross.flow +++ b/h2o-test-integ/tests/flow/BigCross.flow @@ -35,7 +35,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-62983aab-d0ea-4fdb-af4a-3a254f47bd98\",\"training_frame\":\"BigCross1.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"L_BFGS\",\"max_iter\":-1,\"beta_eps\":0,\"standardize\":true,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":true,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-62983aab-d0ea-4fdb-af4a-3a254f47bd98\",\"training_frame\":\"BigCross1.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"solver\":\"L_BFGS\",\"max_iter\":-1,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"gaussian\",\"balance_classes\":true,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", @@ -83,7 +83,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-caad044d-8bd0-41bf-aa46-a426bfb55121\",\"training_frame\":\"BigCross1.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-5512506751270905000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false,\"export_weights_and_biases\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-caad044d-8bd0-41bf-aa46-a426bfb55121\",\"training_frame\":\"BigCross1.hex\",\"dropNA20Cols\":false,\"response_column\":\"C1\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-5512506751270905000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false,\"export_weights_and_biases\":false}" }, { "type": "cs", diff --git a/h2o-test-integ/tests/flow/airlines_all.flow b/h2o-test-integ/tests/flow/airlines_all.flow index 5a114ea821a4..d6f7ed635e59 100644 --- a/h2o-test-integ/tests/flow/airlines_all.flow +++ b/h2o-test-integ/tests/flow/airlines_all.flow @@ -35,7 +35,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-7b8071da-c477-4df9-a252-a7592948c3f4\",\"training_frame\":\"airlines_all.hex\",\"ignored_columns\":[\"Year\",\"Month\",\"DayofMonth\",\"DayOfWeek\",\"DepTime\",\"ArrTime\",\"TailNum\",\"ActualElapsedTime\",\"CRSElapsedTime\",\"AirTime\",\"ArrDelay\",\"DepDelay\",\"Distance\",\"TaxiIn\",\"TaxiOut\",\"Cancelled\",\"CancellationCode\",\"CarrierDelay\",\"WeatherDelay\",\"NASDelay\",\"SecurityDelay\",\"LateAircraftDelay\"],\"dropNA20Cols\":false,\"response_column\":\"IsDepDelayed\",\"solver\":\"ADMM\",\"max_iterations\":50,\"beta_eps\":0,\"standardize\":true,\"family\":\"binomial\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-7b8071da-c477-4df9-a252-a7592948c3f4\",\"training_frame\":\"airlines_all.hex\",\"ignored_columns\":[\"Year\",\"Month\",\"DayofMonth\",\"DayOfWeek\",\"DepTime\",\"ArrTime\",\"TailNum\",\"ActualElapsedTime\",\"CRSElapsedTime\",\"AirTime\",\"ArrDelay\",\"DepDelay\",\"Distance\",\"TaxiIn\",\"TaxiOut\",\"Cancelled\",\"CancellationCode\",\"CarrierDelay\",\"WeatherDelay\",\"NASDelay\",\"SecurityDelay\",\"LateAircraftDelay\"],\"dropNA20Cols\":false,\"response_column\":\"IsDepDelayed\",\"solver\":\"ADMM\",\"max_iterations\":50,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"binomial\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[0.00001],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", @@ -75,7 +75,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-6ef7f4a6-7816-42c0-9f4a-67295738c9a5\",\"training_frame\":\"airlines_all.hex\",\"ignored_columns\":[\"DepTime\",\"ArrTime\",\"Year\",\"Month\",\"DayofMonth\",\"DayOfWeek\",\"TailNum\",\"ActualElapsedTime\",\"CRSElapsedTime\",\"AirTime\",\"ArrDelay\",\"DepDelay\",\"Distance\",\"TaxiIn\",\"TaxiOut\",\"Cancelled\",\"CancellationCode\",\"Diverted\",\"CarrierDelay\",\"WeatherDelay\",\"NASDelay\",\"SecurityDelay\",\"LateAircraftDelay\"],\"dropNA20Cols\":false,\"response_column\":\"IsDepDelayed\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":8286459160428618000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"CrossEntropy\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-6ef7f4a6-7816-42c0-9f4a-67295738c9a5\",\"training_frame\":\"airlines_all.hex\",\"ignored_columns\":[\"DepTime\",\"ArrTime\",\"Year\",\"Month\",\"DayofMonth\",\"DayOfWeek\",\"TailNum\",\"ActualElapsedTime\",\"CRSElapsedTime\",\"AirTime\",\"ArrDelay\",\"DepDelay\",\"Distance\",\"TaxiIn\",\"TaxiOut\",\"Cancelled\",\"CancellationCode\",\"Diverted\",\"CarrierDelay\",\"WeatherDelay\",\"NASDelay\",\"SecurityDelay\",\"LateAircraftDelay\"],\"dropNA20Cols\":false,\"response_column\":\"IsDepDelayed\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":8286459160428618000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"CrossEntropy\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false}" }, { "type": "cs", diff --git a/h2o-test-integ/tests/flow/covtype_data.flow b/h2o-test-integ/tests/flow/covtype_data.flow index e1fc4cd47446..8b9e3fec5a46 100644 --- a/h2o-test-integ/tests/flow/covtype_data.flow +++ b/h2o-test-integ/tests/flow/covtype_data.flow @@ -31,7 +31,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-0e5d9a55-e702-474f-9af8-ba27bad8c5a0\",\"training_frame\":\"covtype.hex\",\"dropNA20Cols\":false,\"response_column\":\"C55\",\"solver\":\"ADMM\",\"max_iter\":-1,\"beta_eps\":0,\"standardize\":true,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-0e5d9a55-e702-474f-9af8-ba27bad8c5a0\",\"training_frame\":\"covtype.hex\",\"dropNA20Cols\":false,\"response_column\":\"C55\",\"solver\":\"ADMM\",\"max_iter\":-1,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"gaussian\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", @@ -79,7 +79,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-4d923821-912f-402b-a379-f33678d2897e\",\"training_frame\":\"covtype.hex\",\"dropNA20Cols\":false,\"response_column\":\"C55\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-1244974365897169000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false,\"export_weights_and_biases\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-4d923821-912f-402b-a379-f33678d2897e\",\"training_frame\":\"covtype.hex\",\"dropNA20Cols\":false,\"response_column\":\"C55\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-1244974365897169000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false,\"export_weights_and_biases\":false}" }, { "type": "cs", diff --git a/h2o-test-integ/tests/flow/prostate.flow b/h2o-test-integ/tests/flow/prostate.flow index c87f75402bcf..cecb61d7f3fe 100644 --- a/h2o-test-integ/tests/flow/prostate.flow +++ b/h2o-test-integ/tests/flow/prostate.flow @@ -31,7 +31,7 @@ }, { "type": "cs", - "input": "buildModel 'glm', {\"destination_key\":\"glm-f3096748-1ed4-402e-8599-4579458ee71a\",\"training_frame\":\"prostate.hex\",\"dropNA20Cols\":false,\"response_column\":\"CAPSULE\",\"solver\":\"ADMM\",\"max_iter\":-1,\"beta_eps\":0,\"standardize\":true,\"family\":\"gaussian\",\"n_folds\":0,\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior1\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" + "input": "buildModel 'glm', {\"destination_key\":\"glm-f3096748-1ed4-402e-8599-4579458ee71a\",\"training_frame\":\"prostate.hex\",\"dropNA20Cols\":false,\"response_column\":\"CAPSULE\",\"solver\":\"ADMM\",\"max_iter\":-1,\"beta_epsilon\":0,\"standardize\":true,\"family\":\"gaussian\",\"balance_classes\":false,\"link\":\"family_default\",\"tweedie_variance_power\":\"NaN\",\"tweedie_link_power\":\"NaN\",\"alpha\":[0.5],\"lambda\":[],\"lambda_search\":false,\"use_all_factor_levels\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"prior\":0,\"nlambdas\":-1,\"lambda_min_ratio\":-1}" }, { "type": "cs", @@ -79,7 +79,7 @@ }, { "type": "cs", - "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-4cfad5f8-b4f0-4d5c-a3d3-e109b2d08a31\",\"training_frame\":\"prostate.hex\",\"dropNA20Cols\":false,\"response_column\":\"CAPSULE\",\"n_folds\":0,\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-5061612826329160000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false,\"export_weights_and_biases\":false}" + "input": "buildModel 'deeplearning', {\"destination_key\":\"deeplearning-4cfad5f8-b4f0-4d5c-a3d3-e109b2d08a31\",\"training_frame\":\"prostate.hex\",\"dropNA20Cols\":false,\"response_column\":\"CAPSULE\",\"activation\":\"Rectifier\",\"hidden\":[50,50],\"epochs\":\"1\",\"variable_importances\":false,\"replicate_training_data\":true,\"balance_classes\":false,\"checkpoint\":\"\",\"use_all_factor_levels\":true,\"train_samples_per_iteration\":-2,\"adaptive_rate\":true,\"rho\":0.99,\"epsilon\":1e-8,\"input_dropout_ratio\":0,\"hidden_dropout_ratios\":[],\"l1\":0,\"l2\":0,\"score_interval\":5,\"score_training_samples\":10000,\"score_validation_samples\":0,\"autoencoder\":false,\"class_sampling_factors\":[],\"max_after_balance_size\":5,\"keep_cross_validation_splits\":false,\"override_with_best_model\":true,\"target_ratio_comm_to_comp\":0.02,\"seed\":-5061612826329160000,\"rate\":0.005,\"rate_annealing\":0.000001,\"rate_decay\":1,\"momentum_start\":0,\"momentum_ramp\":1000000,\"momentum_stable\":0,\"nesterov_accelerated_gradient\":true,\"max_w2\":\"Infinity\",\"initial_weight_distribution\":\"UniformAdaptive\",\"initial_weight_scale\":1,\"loss\":\"Automatic\",\"score_duty_cycle\":0.1,\"classification_stop\":0,\"regression_stop\":0.000001,\"max_hit_ratio_k\":10,\"score_validation_sampling\":\"Uniform\",\"diagnostics\":true,\"fast_mode\":true,\"ignore_const_cols\":true,\"force_load_balance\":true,\"single_node_mode\":false,\"shuffle_training_data\":false,\"missing_values_handling\":\"MeanImputation\",\"quiet_mode\":false,\"max_confusion_matrix_size\":20,\"sparse\":false,\"col_major\":false,\"average_activation\":0,\"sparsity_beta\":0,\"max_categorical_features\":2147483647,\"reproducible\":false,\"export_weights_and_biases\":false}" }, { "type": "cs",