Merge c1fa41b into 58f3e5d

doc/index.rst

@@ -23,21 +23,14 @@ come to know and love.
 Travis CI Build Status
 ======================
 
-.. raw:: html
-
-   <p style="height:22px">
-     <a href="https://travis-ci.org/tgsmith61591/skutil" >
-       <img src="https://travis-ci.org/tgsmith61591/skutil.svg?branch=master"/>
-     </a>
-     <a href="https://coveralls.io/github/tgsmith61591/skutil?branch=master" >
-       <img src="https://coveralls.io/repos/github/tgsmith61591/skutil/badge.svg?branch=master"/>
-     </a>
-     <a href="https://img.shields.io/badge/python-2.7-blue" >
-       <img src="https://img.shields.io/badge/python-2.7-blue.svg"/>
-     </a>
-   </p>
+.. image:: https://travis-ci.org/tgsmith61591/skutil.svg?branch=master
+   :target: https://travis-ci.org/tgsmith61591/skutil
 
+.. image:: https://coveralls.io/repos/github/tgsmith61591/skutil/badge.svg?branch=master
+   :target: https://coveralls.io/github/tgsmith61591/skutil?branch=master
 
+.. image:: https://img.shields.io/badge/python-2.7-blue.svg
+   :target: https://img.shields.io/badge/python-2.7-blue
 
 Helpful Links
 =============

skutil/h2o/grid_search.py

@@ -848,7 +848,7 @@ class H2ORandomizedSearchCV(BaseH2OSearchCV):
 
         n_iter : int, optional (default=10)
             The number of iterations to fit. Note that 
-            n_iter * cv.get_n_splits will be fit. If there
+            ``n_iter * cv.get_n_splits`` will be fit. If there
             are 10 folds and 10 iterations, 100 models (plus
             one) will be fit.
 
@@ -865,7 +865,7 @@ class H2ORandomizedSearchCV(BaseH2OSearchCV):
 
         cv : int or H2OCrossValidator, optional (default=5)
             The number of folds to be fit for cross validation.
-            Note that n_iter * cv.get_n_splits will be fit. If there
+            Note that ``n_iter * cv.get_n_splits`` will be fit. If there
             are 10 folds and 10 iterations, 100 models (plus
             one) will be fit.
 
@@ -984,7 +984,7 @@ class H2OGainsRandomizedSearchCV(H2ORandomizedSearchCV):
 
     n_iter : int, optional (default=10)
         The number of iterations to fit. Note that 
-        n_iter * cv.get_n_splits will be fit. If there
+        ``n_iter * cv.get_n_splits`` will be fit. If there
         are 10 folds and 10 iterations, 100 models (plus
         one) will be fit.
 
@@ -1001,7 +1001,7 @@ class H2OGainsRandomizedSearchCV(H2ORandomizedSearchCV):
 
     cv : int or H2OCrossValidator, optional (default=5)
         The number of folds to be fit for cross validation.
-        Note that n_iter * cv.get_n_splits will be fit. If there
+        Note that ``n_iter * cv.get_n_splits`` will be fit. If there
         are 10 folds and 10 iterations, 100 models (plus
         one) will be fit.
 
@@ -1149,7 +1149,7 @@ def report_scores(self):
     @overrides(BaseH2OSearchCV)
     def score(self, frame):
         """Predict and score on a new frame. Note that this method
-        will not store performance metrics in the report that ```report_score```
+        will not store performance metrics in the report that ``report_score``
         generates.
 
         Parameters

skutil/h2o/metrics.py

@@ -780,7 +780,7 @@ def make_h2o_scorer(score_function, y_true):
     """Make a scoring function from a callable.
     The signature for the callable should resemble:
 
-        ```some_function(y_true, y_pred, y_type=None...)```
+        ``some_function(y_true, y_pred, y_type=None...)``
 
     Parameters
     ----------
@@ -802,9 +802,9 @@ class _H2OScorer(six.with_metaclass(abc.ABCMeta)):
     with H2OFrames. The first two arguments in the scoring function
     signature should resemble the following: 
 
-        ```some_function(y_true, y_pred, y_type=None...)```
+        ``some_function(y_true, y_pred, y_type=None...)``
 
-    Any specific scoring kwargs should be passed to the ```score```
+    Any specific scoring kwargs should be passed to the ``score``
     function in the class instance.
 
     Parameters

skutil/h2o/transform.py

@@ -356,7 +356,9 @@ def _mul(a, b):
 class H2OInteractionTermTransformer(BaseH2OTransformer):
     """A class that will generate interaction terms between selected columns.
     An interaction captures some relationship between two independent variables
-    in the form of In = (xi * xj).
+    in the form of:
+
+    :math:`In = (x_i * x_j)`
 
     Note that the H2OInteractionTermTransformer will only operate on the feature_names,
     and at the transform point will return ALL features plus the newly generated ones.

skutil/linear_model/__init__.py

@@ -1,4 +1,6 @@
 """
 skutil.linear_model will provide the basis for
 Poisson and Tweedy regressions and gradient computations.
-"""
+"""
+
+__all__ = [s for s in dir() if not s.startswith("_")]  # Remove hiddens

skutil/metrics/__init__.py

@@ -6,4 +6,4 @@
 from .pairwise import *
 from ._act import *
 
-# __all__ = [s for s in dir() if not s.startswith("_")]  # Remove hiddens
+__all__ = [s for s in dir() if not s.startswith("_")]  # Remove hiddens

skutil/metrics/_kernel.py → skutil/metrics/kernel.py

@@ -39,7 +39,7 @@ def _prep_X_Y_for_cython(X, Y):
 
 # Cython proxies
 def _hilbert_dot(x, y, scalar=1.0):
-    # return 2 * safe_sparse_dot(x, y) - safe_sparse_dot(x, x.T) - safe_sparse_dot(y, y.T)
+    # return ``2 * safe_sparse_dot(x, y) - safe_sparse_dot(x, x.T) - safe_sparse_dot(y, y.T)``
     x, y = x.astype(np.double, order='C'), y.astype(np.double, order='C')
     return _hilbert_dot_fast(x, y, scalar)
 
@@ -59,6 +59,10 @@ def exponential_kernel(X, Y=None, sigma=1.0):
     power. In the other hand, if underestimated, the function will lack regularization and 
     the decision boundary will be highly sensitive to noise in training data.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = exp( -||x-y|| / 2\\sigma^2 )`
+
     Parameters
     ----------
 
@@ -92,14 +96,18 @@ def exponential_kernel(X, Y=None, sigma=1.0):
 
 
 def gaussian_kernel(X, Y=None, sigma=1.0):
-    """The ``gaussian_kernel`` is closely related to the ``exponential_kernel``. 
+    """The ``gaussian_kernel`` is closely related to the ``exponential_kernel``.
     It is also an ``rbf_kernel``. Note that the adjustable parameter, ``sigma``, 
     plays a major role in the performance of the kernel and should be carefully 
     tuned. If overestimated, the exponential will behave almost linearly and 
     the higher-dimensional projection will start to lose its non-linear 
     power. In the other hand, if underestimated, the function will lack regularization and 
     the decision boundary will be highly sensitive to noise in training data.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = exp( -||x-y||^2 / 2\\sigma^2 )`
+
     Parameters
     ----------
 
@@ -137,6 +145,10 @@ def inverse_multiquadric_kernel(X, Y=None, constant=1.0):
     results in a kernel matrix with full rank (Micchelli, 1986) and thus forms 
     an infinite dimension feature space.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = 1 / sqrt( -||x-y||^2 + c^2 )`
+
     Parameters
     ----------
 
@@ -174,6 +186,10 @@ def laplace_kernel(X, Y=None, sigma=1.0):
     except for being less sensitive for changes in the ``sigma`` parameter. 
     Being equivalent, it is also an ``rbf_kernel``.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = exp( -||x-y|| / \\sigma )`
+
     Parameters
     ----------
 
@@ -212,6 +228,10 @@ def linear_kernel(X, Y=None, constant=0.0):
     Kernel algorithms using a linear kernel are often equivalent to their non-kernel 
     counterparts, i.e. KPCA with a ``linear_kernel`` is the same as standard PCA.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = x^Ty + c`
+
     Parameters
     ----------
 
@@ -249,6 +269,10 @@ def multiquadric_kernel(X, Y=None, constant=0.0):
     as the Rational Quadratic kernel. As is the case with the Sigmoid kernel, 
     it is also an example of an non-positive definite kernel.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = sqrt( -||x-y||^2 + c^2 )`
+
     Parameters
     ----------
 
@@ -289,6 +313,10 @@ def polynomial_kernel(X, Y=None, alpha=1.0, degree=1.0, constant=1.0):
     Adjustable parameters are the slope (``alpha``), the constant term (``constant``), 
     and the polynomial degree (``degree``).
 
+    The kernel is given by:
+
+        :math:`k(x, y) = ( \\alpha x^Ty + c)^d`
+
     Parameters
     ----------
 
@@ -333,6 +361,10 @@ def power_kernel(X, Y=None, degree=1.0):
     It is an example of scale-invariant kernel (Sahbi and Fleuret, 2004) and is 
     also only conditionally positive definite.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = -||x-y||^d`
+
     Parameters
     ----------
 
@@ -374,6 +406,14 @@ def rbf_kernel(X, Y=None, sigma=1.0):
     power. In the other hand, if underestimated, the function will lack regularization and 
     the decision boundary will be highly sensitive to noise in training data.
 
+    The kernel is given by:
+
+        :math:`k(x, y) = exp(- \\gamma * ||x-y||^2)`
+
+    where:
+
+        :math:`\\gamma = 1/( \\sigma ^2)`
+
     Parameters
     ----------
 
@@ -407,9 +447,14 @@ def rbf_kernel(X, Y=None, sigma=1.0):
 
 
 def spline_kernel(X, Y=None):
-    """The ``spline_kernel`` is given as a piece-wise cubic polynomial, 
+    """
+    The ``spline_kernel`` is given as a piece-wise cubic polynomial,
     as derived in the works by Gunn (1998).
 
+   The kernel is given by:
+
+        :math:`k(x, y) = 1 + xy + xy * min(x,y) - (1/2 * (x+y)) * min(x,y)^2 + 1/3 * min(x,y)^3`
+
     Parameters
     ----------
 
@@ -448,7 +493,7 @@ def tanh_kernel(X, Y=None, constant=0.0, alpha=1.0):
 
     The kernel is given by:
 
-        k(x, y) = tanh (alpha x^T y + c)
+        :math:`k(x, y) = tanh (\\alpha x^T y + c)`
 
     It is interesting to note that a SVM model using a sigmoid kernel function is 
     equivalent to a two-layer, perceptron neural network. This kernel was quite popular 

skutil/metrics/pairwise.py

@@ -1,4 +1,4 @@
 import numpy as np
-from ._kernel import *
+from .kernel import *
 from sklearn.utils import check_array, check_X_y
 

skutil/metrics/tests/test_metrics.py

@@ -1,8 +1,8 @@
 from skutil.metrics import *
 import numpy as np
 import timeit
-from skutil.metrics._kernel import (_hilbert_dot,
-                                    _hilbert_matrix)
+from skutil.metrics.kernel import (_hilbert_dot,
+                                   _hilbert_matrix)
 from skutil.metrics import GainsStatisticalReport
 from skutil.utils.tests.utils import assert_fails
 from numpy.testing import (assert_array_equal, assert_almost_equal, assert_array_almost_equal)