import numpy as np import scipy.sparse as sp import warnings from abc import ABCMeta, abstractmethod from . import libsvm, liblinear from . import libsvm_sparse from ..base import BaseEstimator, ClassifierMixin from ..preprocessing import LabelEncoder from ..utils import atleast2d_or_csr, array2d, check_random_state from ..utils import ConvergenceWarning, compute_class_weight, deprecated from ..utils.fixes import unique from ..utils.extmath import safe_sparse_dot LIBSVM_IMPL = ['c_svc', 'nu_svc', 'one_class', 'epsilon_svr', 'nu_svr'] def _one_vs_one_coef(dual_coef, n_support, support_vectors): """Generate primal coefficients from dual coefficients for the one-vs-one multi class LibSVM in the case of a linear kernel.""" # get 1vs1 weights for all n*(n-1) classifiers. # this is somewhat messy. # shape of dual_coef_ is nSV * (n_classes -1) # see docs for details n_class = dual_coef.shape[0] + 1 # XXX we could do preallocation of coef but # would have to take care in the sparse case coef = [] sv_locs = np.cumsum(np.hstack([[0], n_support])) for class1 in xrange(n_class): # SVs for class1: sv1 = support_vectors[sv_locs[class1]:sv_locs[class1 + 1], :] for class2 in xrange(class1 + 1, n_class): # SVs for class1: sv2 = support_vectors[sv_locs[class2]:sv_locs[class2 + 1], :] # dual coef for class1 SVs: alpha1 = dual_coef[class2 - 1, sv_locs[class1]:sv_locs[class1 + 1]] # dual coef for class2 SVs: alpha2 = dual_coef[class1, sv_locs[class2]:sv_locs[class2 + 1]] # build weight for class1 vs class2 coef.append(safe_sparse_dot(alpha1, sv1) + safe_sparse_dot(alpha2, sv2)) return coef class BaseLibSVM(BaseEstimator): """Base class for estimators that use libsvm as backing library This implements support vector machine classification and regression. Parameter documentation is in the derived `SVC` class. """ # see ./classes.py for SVC class. __metaclass__ = ABCMeta _sparse_kernels = ["linear", "poly", "rbf", "sigmoid", "precomputed"] @abstractmethod def __init__(self, impl, kernel, degree, gamma, coef0, tol, C, nu, epsilon, shrinking, probability, cache_size, sparse, class_weight, verbose, max_iter): if not impl in LIBSVM_IMPL: # pragma: no cover raise ValueError("impl should be one of %s, %s was given" % ( LIBSVM_IMPL, impl)) if C is None: # pragma: no cover warnings.warn("Using 'None' for C of BaseLibSVM is deprecated " "since version 0.12, and backward compatibility " "won't be maintained from version 0.14 onward. " "Setting C=1.0.", DeprecationWarning, stacklevel=2) C = 1.0 self.impl = impl self.kernel = kernel self.degree = degree self.gamma = gamma self.coef0 = coef0 self.tol = tol self.C = C self.nu = nu self.epsilon = epsilon self.shrinking = shrinking self.probability = probability self.cache_size = cache_size self.sparse = sparse self.class_weight = class_weight self.verbose = verbose self.max_iter = max_iter @property def _pairwise(self): kernel = self.kernel return kernel == "precomputed" or hasattr(kernel, "__call__") def fit(self, X, y, sample_weight=None): """Fit the SVM model according to the given training data. Parameters ---------- X : {array-like, sparse matrix}, shape = [n_samples, n_features] Training vectors, where n_samples is the number of samples and n_features is the number of features. y : array-like, shape = [n_samples] Target values (class labels in classification, real numbers in regression) sample_weight : array-like, shape = [n_samples], optional Weights applied to individual samples (1. for unweighted). Returns ------- self : object Returns self. Notes ------ If X and y are not C-ordered and contiguous arrays of np.float64 and X is not a scipy.sparse.csr_matrix, X and/or y may be copied. If X is a dense array, then the other methods will not support sparse matrices as input. """ if self.sparse == "auto": self._sparse = sp.isspmatrix(X) and not self._pairwise else: self._sparse = self.sparse if self._sparse and self._pairwise: raise ValueError("Sparse precomputed kernels are not supported. " "Using sparse data and dense kernels is possible " "by not using the ``sparse`` parameter") X = atleast2d_or_csr(X, dtype=np.float64, order='C') if self.impl in ['c_svc', 'nu_svc']: # classification self.classes_, y = unique(y, return_inverse=True) self.class_weight_ = compute_class_weight(self.class_weight, self.classes_, y) else: self.class_weight_ = np.empty(0) if self.impl != "one_class" and len(np.unique(y)) < 2: raise ValueError("The number of classes has to be greater than" " one.") y = np.asarray(y, dtype=np.float64, order='C') sample_weight = np.asarray([] if sample_weight is None else sample_weight, dtype=np.float64) solver_type = LIBSVM_IMPL.index(self.impl) # input validation if solver_type != 2 and X.shape[0] != y.shape[0]: raise ValueError("X and y have incompatible shapes.\n" + "X has %s samples, but y has %s." % (X.shape[0], y.shape[0])) if self.kernel == "precomputed" and X.shape[0] != X.shape[1]: raise ValueError("X.shape[0] should be equal to X.shape[1]") if sample_weight.shape[0] > 0 and sample_weight.shape[0] != X.shape[0]: raise ValueError("sample_weight and X have incompatible shapes:" "%r vs %r\n" "Note: Sparse matrices cannot be indexed w/" "boolean masks (use `indices=True` in CV)." % (sample_weight.shape, X.shape)) if (self.kernel in ['poly', 'rbf']) and (self.gamma == 0): # if custom gamma is not provided ... self._gamma = 1.0 / X.shape[1] else: self._gamma = self.gamma kernel = self.kernel if hasattr(kernel, '__call__'): kernel = 'precomputed' fit = self._sparse_fit if self._sparse else self._dense_fit if self.verbose: # pragma: no cover print '[LibSVM]', fit(X, y, sample_weight, solver_type, kernel) self.shape_fit_ = X.shape # In binary case, we need to flip the sign of coef, intercept and # decision function. Use self._intercept_ internally. self._intercept_ = self.intercept_.copy() if self.impl in ['c_svc', 'nu_svc'] and len(self.classes_) == 2: self.intercept_ *= -1 return self def _warn_from_fit_status(self): assert self.fit_status_ in (0, 1) if self.fit_status_ == 1: warnings.warn('Solver terminated early (max_iter=%i).' ' Consider pre-processing your data with' ' StandardScaler or MinMaxScaler.' % self.max_iter, ConvergenceWarning) def _dense_fit(self, X, y, sample_weight, solver_type, kernel): if hasattr(self.kernel, '__call__'): # you must store a reference to X to compute the kernel in predict # TODO: add keyword copy to copy on demand self.__Xfit = X X = self._compute_kernel(X) if hasattr(self.kernel, '__call__') and X.shape[0] != X.shape[1]: raise ValueError("X.shape[0] should be equal to X.shape[1]") libsvm.set_verbosity_wrap(self.verbose) # we don't pass **self.get_params() to allow subclasses to # add other parameters to __init__ self.support_, self.support_vectors_, self.n_support_, \ self.dual_coef_, self.intercept_, self._label, self.probA_, \ self.probB_, self.fit_status_ = libsvm.fit( X, y, svm_type=solver_type, sample_weight=sample_weight, class_weight=self.class_weight_, kernel=kernel, C=self.C, nu=self.nu, probability=self.probability, degree=self.degree, 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) self._warn_from_fit_status() def _sparse_fit(self, X, y, sample_weight, solver_type, kernel): X.data = np.asarray(X.data, dtype=np.float64, order='C') X.sort_indices() kernel_type = self._sparse_kernels.index(kernel) libsvm_sparse.set_verbosity_wrap(self.verbose) self.support_vectors_, dual_coef_data, self.intercept_, self._label, \ self.n_support_, self.probA_, self.probB_, self.fit_status_ = \ libsvm_sparse.libsvm_sparse_train( X.shape[1], X.data, X.indices, X.indptr, y, solver_type, 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) self._warn_from_fit_status() n_class = len(self._label) - 1 n_SV = self.support_vectors_.shape[0] dual_coef_indices = np.tile(np.arange(n_SV), n_class) dual_coef_indptr = np.arange(0, dual_coef_indices.size + 1, dual_coef_indices.size / n_class) self.dual_coef_ = sp.csr_matrix( (dual_coef_data, dual_coef_indices, dual_coef_indptr), (n_class, n_SV)) def predict(self, X): """Perform regression on samples in X. For an one-class model, +1 or -1 is returned. Parameters ---------- X : {array-like, sparse matrix}, shape = [n_samples, n_features] Returns ------- y_pred : array, shape = [n_samples] """ X = self._validate_for_predict(X) predict = self._sparse_predict if self._sparse else self._dense_predict return predict(X) def _dense_predict(self, X): n_samples, n_features = X.shape X = self._compute_kernel(X) if X.ndim == 1: X = array2d(X, order='C') kernel = self.kernel if hasattr(self.kernel, "__call__"): kernel = 'precomputed' if X.shape[1] != self.shape_fit_[0]: raise ValueError("X.shape[1] = %d should be equal to %d, " "the number of samples at training time" % (X.shape[1], self.shape_fit_[0])) C = 0.0 # C is not useful here svm_type = LIBSVM_IMPL.index(self.impl) return libsvm.predict( X, self.support_, self.support_vectors_, self.n_support_, self.dual_coef_, self._intercept_, self._label, self.probA_, self.probB_, svm_type=svm_type, kernel=kernel, C=C, nu=self.nu, probability=self.probability, degree=self.degree, shrinking=self.shrinking, tol=self.tol, cache_size=self.cache_size, coef0=self.coef0, gamma=self._gamma, epsilon=self.epsilon) def _sparse_predict(self, X): X = sp.csr_matrix(X, dtype=np.float64) kernel = self.kernel if hasattr(kernel, '__call__'): kernel = 'precomputed' kernel_type = self._sparse_kernels.index(kernel) C = 0.0 # C is not useful here return libsvm_sparse.libsvm_sparse_predict( X.data, X.indices, X.indptr, self.support_vectors_.data, self.support_vectors_.indices, self.support_vectors_.indptr, self.dual_coef_.data, self._intercept_, LIBSVM_IMPL.index(self.impl), kernel_type, self.degree, self._gamma, self.coef0, self.tol, C, self.class_weight_, self.nu, self.epsilon, self.shrinking, self.probability, self.n_support_, self._label, self.probA_, self.probB_) def _compute_kernel(self, X): """Return the data transformed by a callable kernel""" if hasattr(self.kernel, '__call__'): # in the case of precomputed kernel given as a function, we # have to compute explicitly the kernel matrix kernel = self.kernel(X, self.__Xfit) if sp.issparse(kernel): kernel = kernel.toarray() X = np.asarray(kernel, dtype=np.float64, order='C') return X def decision_function(self, X): """Distance of the samples X to the separating hyperplane. Parameters ---------- X : array-like, shape = [n_samples, n_features] Returns ------- X : array-like, shape = [n_samples, n_class * (n_class-1) / 2] Returns the decision function of the sample for each class in the model. """ if self._sparse: raise NotImplementedError("Decision_function not supported for" " sparse SVM.") X = self._validate_for_predict(X) C = 0.0 # C is not useful here kernel = self.kernel if hasattr(kernel, '__call__'): kernel = 'precomputed' dec_func = libsvm.decision_function( X, self.support_, self.support_vectors_, self.n_support_, self.dual_coef_, self._intercept_, self._label, self.probA_, self.probB_, svm_type=LIBSVM_IMPL.index(self.impl), kernel=kernel, C=C, nu=self.nu, probability=self.probability, degree=self.degree, shrinking=self.shrinking, tol=self.tol, cache_size=self.cache_size, coef0=self.coef0, gamma=self._gamma, epsilon=self.epsilon) # In binary case, we need to flip the sign of coef, intercept and # decision function. if self.impl in ['c_svc', 'nu_svc'] and len(self.classes_) == 2: return -dec_func return dec_func def _validate_for_predict(self, X): X = atleast2d_or_csr(X, dtype=np.float64, order="C") if self._sparse and not sp.isspmatrix(X): X = sp.csr_matrix(X) if self._sparse: X.sort_indices() if (sp.issparse(X) and not self._sparse and not hasattr(self.kernel, '__call__')): raise ValueError( "cannot use sparse input in %r trained on dense data" % type(self).__name__) n_samples, n_features = X.shape if self.kernel == "precomputed": if X.shape[1] != self.shape_fit_[0]: raise ValueError("X.shape[1] = %d should be equal to %d, " "the number of samples at training time" % (X.shape[1], self.shape_fit_[0])) elif n_features != self.shape_fit_[1]: raise ValueError("X.shape[1] = %d should be equal to %d, " "the number of features at training time" % (n_features, self.shape_fit_[1])) return X @property def coef_(self): if self.kernel != 'linear': raise ValueError('coef_ is only available when using a ' 'linear kernel') if self.dual_coef_.shape[0] == 1: # binary classifier coef = -safe_sparse_dot(self.dual_coef_, self.support_vectors_) else: # 1vs1 classifier coef = _one_vs_one_coef(self.dual_coef_, self.n_support_, self.support_vectors_) if sp.issparse(coef[0]): coef = sp.vstack(coef).tocsr() else: coef = np.vstack(coef) # coef_ being a read-only property it's better to mark the value as # immutable to avoid hiding potential bugs for the unsuspecting user if sp.issparse(coef): # sparse matrix do not have global flags coef.data.flags.writeable = False else: # regular dense array coef.flags.writeable = False return coef class BaseSVC(BaseLibSVM, ClassifierMixin): """ABC for LibSVM-based classifiers.""" def predict(self, X): """Perform classification on samples in X. For an one-class model, +1 or -1 is returned. Parameters ---------- X : {array-like, sparse matrix}, shape = [n_samples, n_features] Returns ------- y_pred : array, shape = [n_samples] Class labels for samples in X. """ y = super(BaseSVC, self).predict(X) return self.classes_.take(y.astype(np.int)) def predict_proba(self, X): """Compute probabilities of possible outcomes for samples in X. The model need to have probability information computed at training time: fit with attribute `probability` set to True. Parameters ---------- X : array-like, shape = [n_samples, n_features] Returns ------- X : array-like, shape = [n_samples, n_classes] Returns the probability of the sample for each class in the model, where classes are ordered by arithmetical order. Notes ----- The probability model is created using cross validation, so the results can be slightly different than those obtained by predict. Also, it will produce meaningless results on very small datasets. """ if not self.probability: raise NotImplementedError( "probability estimates must be enabled to use this method") if self.impl not in ('c_svc', 'nu_svc'): raise NotImplementedError("predict_proba only implemented for SVC " "and NuSVC") X = self._validate_for_predict(X) pred_proba = (self._sparse_predict_proba if self._sparse else self._dense_predict_proba) return pred_proba(X) def predict_log_proba(self, X): """Compute log probabilities of possible outcomes for samples in X. The model need to have probability information computed at training time: fit with attribute `probability` set to True. Parameters ---------- X : array-like, shape = [n_samples, n_features] Returns ------- X : array-like, shape = [n_samples, n_classes] Returns the log-probabilities of the sample for each class in the model, where classes are ordered by arithmetical order. Notes ----- The probability model is created using cross validation, so the results can be slightly different than those obtained by predict. Also, it will produce meaningless results on very small datasets. """ return np.log(self.predict_proba(X)) def _dense_predict_proba(self, X): X = self._compute_kernel(X) C = 0.0 # C is not useful here kernel = self.kernel if hasattr(kernel, '__call__'): kernel = 'precomputed' svm_type = LIBSVM_IMPL.index(self.impl) pprob = libsvm.predict_proba( X, self.support_, self.support_vectors_, self.n_support_, self.dual_coef_, self._intercept_, self._label, self.probA_, self.probB_, svm_type=svm_type, kernel=kernel, C=C, nu=self.nu, probability=self.probability, degree=self.degree, shrinking=self.shrinking, tol=self.tol, cache_size=self.cache_size, coef0=self.coef0, gamma=self._gamma, epsilon=self.epsilon) return pprob def _sparse_predict_proba(self, X): X.data = np.asarray(X.data, dtype=np.float64, order='C') kernel = self.kernel if hasattr(kernel, '__call__'): kernel = 'precomputed' kernel_type = self._sparse_kernels.index(kernel) return libsvm_sparse.libsvm_sparse_predict_proba( X.data, X.indices, X.indptr, self.support_vectors_.data, self.support_vectors_.indices, self.support_vectors_.indptr, self.dual_coef_.data, self._intercept_, LIBSVM_IMPL.index(self.impl), kernel_type, self.degree, self._gamma, self.coef0, self.tol, self.C, self.class_weight_, self.nu, self.epsilon, self.shrinking, self.probability, self.n_support_, self._label, self.probA_, self.probB_) @property @deprecated("The ``label_`` attribute has been renamed to ``classes_`` " "for consistency and will be removed in 0.15.") def label_(self): return self.classes_ class BaseLibLinear(BaseEstimator): """Base for classes binding liblinear (dense and sparse versions)""" _solver_type_dict = { 'PL2_LLR_D0': 0, # L2 penalty, logistic regression 'PL2_LL2_D1': 1, # L2 penalty, L2 loss, dual form 'PL2_LL2_D0': 2, # L2 penalty, L2 loss, primal form 'PL2_LL1_D1': 3, # L2 penalty, L1 Loss, dual form 'MC_SVC': 4, # Multi-class Support Vector Classification 'PL1_LL2_D0': 5, # L1 penalty, L2 Loss, primal form 'PL1_LLR_D0': 6, # L1 penalty, logistic regression 'PL2_LLR_D1': 7, # L2 penalty, logistic regression, dual form } def __init__(self, penalty='l2', loss='l2', dual=True, tol=1e-4, C=1.0, multi_class='ovr', fit_intercept=True, intercept_scaling=1, class_weight=None, verbose=0, random_state=None): if C is None: # pragma: no cover warnings.warn("Using 'None' for C of BaseLibLinear is deprecated " "since version 0.12, and backward compatibility " "won't be maintained from version 0.14 onward. " "Setting C=1.0.", DeprecationWarning, stacklevel=2) C = 1.0 self.penalty = penalty self.loss = loss self.dual = dual self.tol = tol self.C = C self.fit_intercept = fit_intercept self.intercept_scaling = intercept_scaling self.multi_class = multi_class self.class_weight = class_weight self.verbose = verbose self.random_state = random_state # Check that the arguments given are valid: self._get_solver_type() def _get_solver_type(self): """Find the liblinear magic number for the solver. This number depends on the values of the following attributes: - multi_class - penalty - loss - dual """ if self.multi_class == 'crammer_singer': solver_type = 'MC_SVC' else: if self.multi_class != 'ovr': raise ValueError("`multi_class` must be one of `ovr`, " "`crammer_singer`") solver_type = "P%s_L%s_D%d" % ( self.penalty.upper(), self.loss.upper(), int(self.dual)) if not solver_type in self._solver_type_dict: if self.penalty.upper() == 'L1' and self.loss.upper() == 'L1': error_string = ("The combination of penalty='l1' " "and loss='l1' is not supported.") elif self.penalty.upper() == 'L2' and self.loss.upper() == 'L1': # this has to be in primal error_string = ("penalty='l2' and loss='l1' is " "only supported when dual='true'.") else: # only PL1 in dual remains error_string = ("penalty='l1' is only supported " "when dual='false'.") raise ValueError('Not supported set of arguments: ' + error_string) return self._solver_type_dict[solver_type] def fit(self, X, y): """Fit the model according to the given training data. Parameters ---------- X : {array-like, sparse matrix}, shape = [n_samples, n_features] Training vector, where n_samples in the number of samples and n_features is the number of features. y : array-like, shape = [n_samples] Target vector relative to X class_weight : {dict, 'auto'}, optional Weights associated with classes. If not given, all classes are supposed to have weight one. Returns ------- self : object Returns self. """ self._enc = LabelEncoder() y = self._enc.fit_transform(y) if len(self.classes_) < 2: raise ValueError("The number of classes has to be greater than" " one.") X = atleast2d_or_csr(X, dtype=np.float64, order="C") y = np.asarray(y, dtype=np.float64).ravel() self.class_weight_ = compute_class_weight(self.class_weight, self.classes_, y) if X.shape[0] != y.shape[0]: raise ValueError("X and y have incompatible shapes.\n" "X has %s samples, but y has %s." % (X.shape[0], y.shape[0])) liblinear.set_verbosity_wrap(self.verbose) if sp.isspmatrix(X): train = liblinear.csr_train_wrap else: train = liblinear.train_wrap rnd = check_random_state(self.random_state) if self.verbose: print '[LibLinear]', self.raw_coef_ = train(X, y, self._get_solver_type(), self.tol, self._get_bias(), self.C, self.class_weight_, # seed for srand in range [0..INT_MAX); # due to limitations in Numpy on 32-bit # platforms, we can't get to the UINT_MAX # limit that srand supports rnd.randint(np.iinfo('i').max)) if self.fit_intercept: self.coef_ = self.raw_coef_[:, :-1] self.intercept_ = self.intercept_scaling * self.raw_coef_[:, -1] else: self.coef_ = self.raw_coef_ self.intercept_ = 0. return self @property def classes_(self): return self._enc.classes_ @property @deprecated("The ``label_`` attribute has been renamed to ``classes_`` " "for consistency and will be removed in 0.15.") def label_(self): return self._enc.classes_ def _get_bias(self): if self.fit_intercept: return self.intercept_scaling else: return -1.0 libsvm.set_verbosity_wrap(0) libsvm_sparse.set_verbosity_wrap(0) liblinear.set_verbosity_wrap(0)