algorithms.yml

Supervised:
  defaults:
    metrics: classification
    preprocessors:
      - MM
  AB:
    base:        !!python/name:sklearn.ensemble.AdaBoostClassifier
    description: Adaptive Boosting
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostClassifier.html
    parameters:
      static:
        algorithm:     SAMME.R
        learning_rate: 1.
        n_estimators:  50
        random_state:  42
  BN:
    base:        !!python/name:pbox.core.model.algorithm.weka.BayesNet  # based on weka.classifiers.bayes.BayesNet
    boolean:     true
    description: Bayesian Network
    link:        https://weka.sourceforge.io/doc.stable/weka/classifiers/bayes/BayesNet.html
    parameters:
      static:
        E: weka.classifiers.bayes.net.estimate.SimpleEstimator  # estimator algorithm
        Q: weka.classifiers.bayes.net.search.SearchAlgorithm    # search algorithm
  BNB:
    base:        !!python/name:sklearn.naive_bayes.BernoulliNB
    boolean:     true
    description: Bernoulli Naive Bayes
    link:        http://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.BernoulliNB.html
    multiclass:  false
    parameters:
      static:
        alpha:       1.
        binarize:    0.
        fit_prior:   true
        class_prior: null
  D:
    base:        !!python/name:pbox.core.model.algorithm.weka.Decorate  # based on weka.classifiers.meta.Decorate
    description: Decorate
    link:        https://weka.sourceforge.io/doc.stable/weka/classifiers/meta/Decorate.html
    parameters:
      static:
        E: 15                          # desired size of ensemble
        R: 1.0                         # factor that determines number of artificial examples to generate
        S: 42                          # random number seed
        I: 50                          # number of iterations
        W: weka.classifiers.trees.J48  # full name of base classifier
  DL85:
    base:        !!python/name:pydl85.DL85Classifier
    description: DL8.5 (optimal decision tree)
    link:        https://dl85.readthedocs.io/en/latest/
    multiclass:  false
    parameters:
      static:
        time_limit: 60
      cv:
        max_depth: !!python/object/new:range [3, 6]
  DT:
    base:        !!python/name:sklearn.tree.DecisionTreeClassifier
    description: Decision Tree
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html
    parameters:
      static:
        class_weight:   null
        criterion:      entropy
        max_features:   null
        max_leaf_nodes: null
        splitter:       best
        random_state:   42
      cv:
        max_depth: !!python/object/new:range [2, 13, 2]
  GB:
    base:        !!python/name:sklearn.ensemble.GradientBoostingClassifier
    description: Gradient Boosting
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html
    parameters:
      static:
        n_estimators: 30
        random_state: 42
      cv:
        max_depth: !!python/object/new:range [3, 8]
  GNB:
    base:        !!python/name:sklearn.naive_bayes.GaussianNB
    description: Gaussian Naive Bayes
    link:        http://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.GaussianNB.html
    parameters:
      static:
        priors: null
  J48:
    base:        !!python/name:pbox.core.model.algorithm.weka.J48  # based on weka.classifiers.trees.J48
    description: Decision Tree (J48)
    link:        https://weka.sourceforge.io/doc.stable/weka/classifiers/trees/J48.html
    parameters:
      static:
        C: 0.25  # confidence threshold for pruning
        M: 2     # minimum number of instances per leaf
        Q: 42    # random number seed
  kNN:
    base:        !!python/name:sklearn.neighbors.KNeighborsClassifier
    description: k-Nearest Neighbors
    link:        http://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html
    parameters:
      static:
        leaf_size:   30
        metric:      minkowski
      cv:
        n_neighbors: !!python/object/new:range [1, 10, 2]
        p:           [1, 2]                        # euclidian distance
        weights:     ["distance", "uniform"]
    preprocessors:
      - Std
      - PCA
  LP:
    base:        !!python/name:sklearn.linear_model.Perceptron
    description: Linear Perceptron
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Perceptron.html
    parameters:
      static:
        max_iter:     1000
        random_state: 42
  LR:
    base:        !!python/name:pbox.core.model.algorithm.weka.Logistic  # based on weka.classifiers.functions.Logistic
    description: Logistic Regression
    link:        https://weka.sourceforge.io/doc.stable/weka/classifiers/functions/Logistic.html
    parameters:
      static:
        M: -1  # maximum number of iterations (-1: until convergence)
  LSVC:
    base:        !!python/name:sklearn.svm.LinearSVC
    description: Linear Support Vector Classification
    link:       http://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVC.html
    parameters:
      static:
        dual:         true
        loss:         squared_hinge
        penalty:      l2
        random_state: 42
        tol:          !!float 1e-4
      cv:
        C: !!python/object/new:numpy.logspace [-1, 4, 6]
  MLP:
    base:        !!python/name:sklearn.neural_network.MLPClassifier
    description: Multi-Layer Perceptron
    link:        http://scikit-learn.org/stable/modules/generated/sklearn.neural_network.MLPClassifier.html
    parameters:
      static:
        activation:         relu
        alpha:              !!float 1e-4
        hidden_layer_sizes: !!python/tuple [10, 2]
        random_state:       42
        solver:             adam
  MNB:
    base:        !!python/name:sklearn.naive_bayes.MultinomialNB
    description: Multinomial Naive Bayes
    link:        http://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html
    parameters:
      static:
        alpha:       1.
        fit_prior:   true
        class_prior: null
  RF:
    base:        !!python/name:sklearn.ensemble.RandomForestClassifier
    description: Random Forest
    link:        http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html
    parameters:
      static:
        max_depth:    null
        max_features: sqrt
        random_state: 42
      cv:
        criterion:    ["entropy", "gini"]
        max_depth:    !!python/object/new:range [4, 11]
        n_estimators: [10, 20, 30, 40]
  SVM:
    base:        !!python/name:sklearn.svm.SVC
    description: Support Vector Machine
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html
    parameters:
      static:
        C:            1.
        degree:       3
        gamma:        auto
        kernel:       rbf
        probability:  true
        random_state: 42
      cv:
        C:     !!python/object/new:numpy.logspace [4, 6, 3]
        gamma: !!python/object/new:numpy.logspace [-3, -1, 3]

Unsupervised:
  defaults:
    metrics:
      - classification
      - clustering
    preprocessors:
      - MM
  AC:
    base:        !!python/name:sklearn.cluster.AgglomerativeClustering
    description: Agglomerative Clustering
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.AgglomerativeClustering.html
    parameters:
      n_clusters:         auto
      distance_threshold: null
      linkage:            ward
  AP:
    base:        !!python/name:sklearn.cluster.AffinityPropagation
    description: Affinity Propagation clustering
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.AffinityPropagation.html
    parameters:
      static:
        damping:          0.5
        max_iter:         200
        convergence_iter: 15
        affinity :        euclidean
  Birch:
    base:        !!python/name:sklearn.cluster.Birch
    description: BIRCH clustering
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.Birch.html
    parameters:
      static:
        threshold:        0.5
        branching_factor: 50
        n_clusters:       auto
  DBSCAN:
    base:        !!python/name:sklearn.cluster.DBSCAN
    description: Density-Based Spatial Clustering of Applications with Noise
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html
    parameters:
      static:
        eps:          0.5
        min_samples:  5
        leaf_size:    30
        algorithm :   auto 
  KMeans:
    base:        !!python/name:sklearn.cluster.KMeans
    description: K-Means clustering
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html
    parameters:
      static:
        n_clusters: auto
        n_init:     10
        max_iter:   300
        tol:        !!float 1e-4
        algorithm:  lloyd
  MBKMeans:
    base:        !!python/name:sklearn.cluster.MiniBatchKMeans
    description: Mini-Batch K-Means clustering
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.MiniBatchKMeans.html
    parameters:
      static:
        n_clusters:         auto
        batch_size:         1024
        max_iter:           300
        tol:                0.0
        max_no_improvement: 10
        n_init:             3
        reassignment_ratio: 0.1
  MS:
    base:        !!python/name:sklearn.cluster.MeanShift
    description: Mean Shift
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.MeanShift.html
    parameters:
      static:
        bandwidth: null
        max_iter:  300
  OPTICS:
    base:        !!python/name:sklearn.cluster.OPTICS
    description: Ordering Points To Identify the Clustering Structure
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.OPTICS.html
    parameters:
      static:
        min_samples:      5
        p:                2
        min_cluster_size: null
        cluster_method :  xi 
        algorithm:        auto 
  SC:
    base:        !!python/name:sklearn.cluster.SpectralClustering
    description: Spectral Clustering
    link:        https://scikit-learn.org/stable/modules/generated/sklearn.cluster.SpectralClustering.html
    parameters:
      static:
        n_clusters:   auto
        n_components: 8
        n_init:       10
        gamma:        1.0
        n_neighbors:  10
        degree:       3
        coef0:        0
        eigen_solver: null