FIX support random state in libsvm

Author: vene

sklearn/svm/base.py

@@ -68,7 +68,7 @@ class BaseLibSVM(six.with_metaclass(ABCMeta, BaseEstimator)):
     @abstractmethod
     def __init__(self, impl, kernel, degree, gamma, coef0,
                  tol, C, nu, epsilon, shrinking, probability, cache_size,
-                 class_weight, verbose, max_iter):
+                 class_weight, verbose, max_iter, random_state):
 
         if not impl in LIBSVM_IMPL:  # pragma: no cover
             raise ValueError("impl should be one of %s, %s was given" % (
@@ -89,6 +89,7 @@ def __init__(self, impl, kernel, degree, gamma, coef0,
         self.class_weight = class_weight
         self.verbose = verbose
         self.max_iter = max_iter
+        self.random_state = random_state
 
     @property
     def _pairwise(self):
@@ -126,6 +127,8 @@ def fit(self, X, y, sample_weight=None):
         matrices as input.
         """
 
+        rnd = check_random_state(self.random_state)
+
         self._sparse = sp.isspmatrix(X) and not self._pairwise
 
         if self._sparse and self._pairwise:
@@ -170,7 +173,10 @@ def fit(self, X, y, sample_weight=None):
         fit = self._sparse_fit if self._sparse else self._dense_fit
         if self.verbose:  # pragma: no cover
             print('[LibSVM]', end='')
-        fit(X, y, sample_weight, solver_type, kernel)
+
+        seed = rnd.randint(np.iinfo('i').max)
+        fit(X, y, sample_weight, solver_type, kernel, random_seed=seed)
+        # see comment on the other call to np.iinfo in this file
 
         self.shape_fit_ = X.shape
 
@@ -199,7 +205,8 @@ def _warn_from_fit_status(self):
                           ' StandardScaler or MinMaxScaler.'
                           % self.max_iter, ConvergenceWarning)
 
-    def _dense_fit(self, X, y, sample_weight, solver_type, kernel):
+    def _dense_fit(self, X, y, sample_weight, solver_type, kernel,
+                   random_seed):
         if callable(self.kernel):
             # you must store a reference to X to compute the kernel in predict
             # TODO: add keyword copy to copy on demand
@@ -223,11 +230,12 @@ def _dense_fit(self, X, y, sample_weight, solver_type, kernel):
                 shrinking=self.shrinking, tol=self.tol,
                 cache_size=self.cache_size, coef0=self.coef0,
                 gamma=self._gamma, epsilon=self.epsilon,
-                max_iter=self.max_iter)
+                max_iter=self.max_iter, random_seed=random_seed)
 
         self._warn_from_fit_status()
 
-    def _sparse_fit(self, X, y, sample_weight, solver_type, kernel):
+    def _sparse_fit(self, X, y, sample_weight, solver_type, kernel,
+                    random_seed):
         X.data = np.asarray(X.data, dtype=np.float64, order='C')
         X.sort_indices()
 
@@ -243,7 +251,8 @@ def _sparse_fit(self, X, y, sample_weight, solver_type, kernel):
                 kernel_type, self.degree, self._gamma, self.coef0, self.tol,
                 self.C, self.class_weight_,
                 sample_weight, self.nu, self.cache_size, self.epsilon,
-                int(self.shrinking), int(self.probability), self.max_iter)
+                int(self.shrinking), int(self.probability), self.max_iter,
+                random_seed)
 
         self._warn_from_fit_status()
 

sklearn/svm/classes.py

@@ -76,7 +76,7 @@ class frequencies.
         per-process runtime setting in liblinear that, if enabled, may not work
         properly in a multithreaded context.
 
-    random_state: int seed, RandomState instance, or None (default)
+    random_state : int seed, RandomState instance, or None (default)
         The seed of the pseudo random number generator to use when
         shuffling the data.
 
@@ -206,6 +206,10 @@ class frequencies.
     max_iter : int, optional (default=-1)
         Hard limit on iterations within solver, or -1 for no limit.
 
+    random_state : int seed, RandomState instance, or None (default)
+        The seed of the pseudo random number generator to use when
+        shuffling the data for probability estimation.
+
     Attributes
     ----------
     `support_` : array-like, shape = [n_SV]
@@ -263,11 +267,12 @@ class frequencies.
     def __init__(self, C=1.0, kernel='rbf', degree=3, gamma=0.0,
                  coef0=0.0, shrinking=True, probability=False,
                  tol=1e-3, cache_size=200, class_weight=None,
-                 verbose=False, max_iter=-1):
+                 verbose=False, max_iter=-1, random_state=None):
 
         super(SVC, self).__init__(
             'c_svc', kernel, degree, gamma, coef0, tol, C, 0., 0., shrinking,
-            probability, cache_size, class_weight, verbose, max_iter)
+            probability, cache_size, class_weight, verbose, max_iter,
+            random_state)
 
 
 class NuSVC(BaseSVC):
@@ -325,6 +330,10 @@ class NuSVC(BaseSVC):
     max_iter : int, optional (default=-1)
         Hard limit on iterations within solver, or -1 for no limit.
 
+    random_state : int seed, RandomState instance, or None (default)
+        The seed of the pseudo random number generator to use when
+        shuffling the data for probability estimation.
+
     Attributes
     ----------
     `support_` : array-like, shape = [n_SV]
@@ -379,11 +388,12 @@ class NuSVC(BaseSVC):
 
     def __init__(self, nu=0.5, kernel='rbf', degree=3, gamma=0.0,
                  coef0=0.0, shrinking=True, probability=False,
-                 tol=1e-3, cache_size=200, verbose=False, max_iter=-1):
+                 tol=1e-3, cache_size=200, verbose=False, max_iter=-1,
+                 random_state=None):
 
         super(NuSVC, self).__init__(
             'nu_svc', kernel, degree, gamma, coef0, tol, 0., nu, 0., shrinking,
-            probability, cache_size, None, verbose, max_iter)
+            probability, cache_size, None, verbose, max_iter, random_state)
 
 
 class SVR(BaseLibSVM, RegressorMixin):
@@ -444,6 +454,10 @@ class SVR(BaseLibSVM, RegressorMixin):
     max_iter : int, optional (default=-1)
         Hard limit on iterations within solver, or -1 for no limit.
 
+    random_state : int seed, RandomState instance, or None (default)
+        The seed of the pseudo random number generator to use when
+        shuffling the data for probability estimaton.
+
     Attributes
     ----------
     `support_` : array-like, shape = [n_SV]
@@ -488,12 +502,13 @@ class SVR(BaseLibSVM, RegressorMixin):
     """
     def __init__(self, kernel='rbf', degree=3, gamma=0.0, coef0=0.0, tol=1e-3,
                  C=1.0, epsilon=0.1, shrinking=True, probability=False,
-                 cache_size=200, verbose=False, max_iter=-1):
+                 cache_size=200, verbose=False, max_iter=-1,
+                 random_state=None):
 
         super(SVR, self).__init__(
             'epsilon_svr', kernel, degree, gamma, coef0, tol, C, 0., epsilon,
             shrinking, probability, cache_size, None, verbose,
-            max_iter)
+            max_iter, random_state)
 
 
 class NuSVR(BaseLibSVM, RegressorMixin):
@@ -555,6 +570,10 @@ class NuSVR(BaseLibSVM, RegressorMixin):
     max_iter : int, optional (default=-1)
         Hard limit on iterations within solver, or -1 for no limit.
 
+    random_state : int seed, RandomState instance, or None (default)
+        The seed of the pseudo random number generator to use when
+        shuffling the data for probability estimation.
+
     Attributes
     ----------
     `support_` : array-like, shape = [n_SV]
@@ -603,11 +622,11 @@ class NuSVR(BaseLibSVM, RegressorMixin):
     def __init__(self, nu=0.5, C=1.0, kernel='rbf', degree=3,
                  gamma=0.0, coef0=0.0, shrinking=True,
                  probability=False, tol=1e-3, cache_size=200,
-                 verbose=False, max_iter=-1):
+                 verbose=False, max_iter=-1, random_state=None):
 
         super(NuSVR, self).__init__(
             'nu_svr', kernel, degree, gamma, coef0, tol, C, nu, 0., shrinking,
-            probability, cache_size, None, verbose, max_iter)
+            probability, cache_size, None, verbose, max_iter, random_state)
 
 
 class OneClassSVM(BaseLibSVM):
@@ -660,6 +679,10 @@ class OneClassSVM(BaseLibSVM):
     max_iter : int, optional (default=-1)
         Hard limit on iterations within solver, or -1 for no limit.
 
+    random_state : int seed, RandomState instance, or None (default)
+        The seed of the pseudo random number generator to use when
+        shuffling the data for probability estimation.
+
     Attributes
     ----------
     `support_` : array-like, shape = [n_SV]
@@ -684,11 +707,12 @@ class OneClassSVM(BaseLibSVM):
     """
     def __init__(self, kernel='rbf', degree=3, gamma=0.0, coef0=0.0, tol=1e-3,
                  nu=0.5, shrinking=True, cache_size=200, verbose=False,
-                 max_iter=-1):
+                 max_iter=-1, random_state=None):
 
         super(OneClassSVM, self).__init__(
             'one_class', kernel, degree, gamma, coef0, tol, 0., nu, 0.,
-            shrinking, False, cache_size, None, verbose, max_iter)
+            shrinking, False, cache_size, None, verbose, max_iter,
+            random_state)
 
     def fit(self, X, sample_weight=None, **params):
         """