a j]@sPddlZddlmZmZddlmZmZddlZddl m Z ddl m Z mZmZddlmZmZddlmZddlmZmZmZmZdd lmZmZdd lmZdd lm Z dd l!m"Z"m#Z#dd l$m%Z%m&Z&m'Z'm(Z(m)Z)m*Z*ddl+m,Z-ddl+m.Z/ddl+m0Z1gdZ2ddZ3Gddde edZ4Gdddee4edZ5ddZ6d!dd Z7dS)"N)ABCMetaabstractmethod)IntegralReal) BaseEstimatorClassifierMixin _fit_context)ConvergenceWarningNotFittedError) LabelEncoder) check_arraycheck_random_state column_or_1dcompute_class_weight)Interval StrOptions)safe_sparse_dot) available_if)_ovr_decision_functioncheck_classification_targets)_check_large_sparse_check_sample_weight _num_samplescheck_consistent_lengthcheck_is_fitted validate_data) _liblinear)_libsvm)_libsvm_sparse)c_svcnu_svcZ one_classZ epsilon_svrZnu_svrc Cs|jdd}g}ttdg|g}t|D]}|||||dddf}t|d|D]z}|||||dddf} ||d||||df} ||||||df} |t| |t| | qbq0|S)zGenerate primal coefficients from dual coefficients for the one-vs-one multi class LibSVM in the case of a linear kernel.rrN)shapenpZcumsumZhstackrangeappendr) dual_coefZ n_supportZsupport_vectorsn_classcoefZsv_locsZclass1Zsv1Zclass2Zsv2Zalpha1Zalpha2r*7E:\soccer\ml_env\Lib\site-packages\sklearn/svm/_base.py_one_vs_one_coef%s     r,cseZdZUdZehdegeeddddgeddheed dddgeeddd dgeed dd dgeed dd dgeed d d dgeed dddgd gd geeddd dgedhe dgdgeeddddgdgdZ e e d<gdZ e ddZfddZeddd;ddZddZdd Zd!d"Zd#d$Zd%d&Zd'd(Zd)d*Zd+d,Zd-d.Zd/d0Zd1d2Zd3d4Zed5d6Zd7d8Z ed9d:Z!Z"S)< BaseLibSVMzBase 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. >linearpoly precomputedsigmoidrbfrNleft)closedscaleautoZneitherright?booleanZbalancedverbose random_statekerneldegreegammacoef0tolCnuepsilon shrinking probability cache_size class_weightr;max_iterr=_parameter_constraints)r.r/r2r1r0cCsz|jtvrtdt|jf||_||_||_||_||_||_||_ ||_ | |_ | |_ | |_ | |_| |_||_||_dS)Nz&impl should be one of %s, %s was given)_impl LIBSVM_IMPL ValueErrorr?r@rArBrCrDrErFrGrHrIrJr;rKr=)selfr?r@rArBrCrDrErFrGrHrIrJr;rKr=r*r*r+__init__js&  zBaseLibSVM.__init__cs*t}|jdk|j_|jdk|j_|SNr0)super__sklearn_tags__r? input_tagspairwisesparserPtags __class__r*r+rTs zBaseLibSVM.__sklearn_tags__T)Zprefer_skip_nested_validationc Cs$t|j}t|}|r*|jdkr*td|o8t|j |_t|jrRt||nt |||t j dddd\}}| |}t j |durgn|t j d}t|j}t|}|d kr||jd krtd d ||jd f|jdkr||jd krtd|jd |jd |jd d krD|jd |krDtd|j|jft|jrTdn|j}|dkrld|_nt|jtr|jdkr|r|||d n|} | d krd|jd | nd|_n|jdkrd|jd |_nt|jtr|j|_|jr|jn|j} |jr.t ddd|!t "dj#} | |||||| dt$|drf|jn|f|_%|j&'|_(|j)|_*|jdvrt+|j,d kr|j&d9_&|j) |_)|jr|j*j-n|j*} t .|j(/} t .| /}| r|std|jdvr|j0|_1n |j02|_1|S)aFit the SVM model according to the given training data. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) or (n_samples, n_samples) Training vectors, where `n_samples` is the number of samples and `n_features` is the number of features. For kernel="precomputed", the expected shape of X is (n_samples, n_samples). y : array-like of shape (n_samples,) Target values (class labels in classification, real numbers in regression). sample_weight : array-like of shape (n_samples,), default=None Per-sample weights. Rescale C per sample. Higher weights force the classifier to put more emphasis on these points. Returns ------- self : object Fitted estimator. 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. r0z-Sparse precomputed kernels are not supported.rDcsrF)dtypeorder accept_sparseaccept_large_sparseNr]rrz"X and y have incompatible shapes. zX has %s samples, but y has %s.rzDPrecomputed matrix must be a square matrix. Input is a {}x{} matrix.zsample_weight and X have incompatible shapes: %r vs %r Note: Sparse matrices cannot be indexed w/boolean masks (use `indices=True` in CV).r7r5r9r6z[LibSVM]endi) random_seedr#r!r"r<zxThe dual coefficients or intercepts are not finite. The input data may contain large values and need to be preprocessed.)3rr=spissparser? TypeErrorcallable_sparserrr$float64_validate_targetsasarrayrNindexrMrr#rOformat_gamma isinstancerAstrmultiplymeanvarr _sparse_fit _dense_fitr;printrandintiinfomaxhasattr shape_fit_ intercept_copy _intercept_ dual_coef_ _dual_coef_lenclasses_dataisfiniteall _num_itern_iter_item)rPXy sample_weightrndrW solver_typeZ n_samplesr?ZX_varfitseedr'Zintercept_finitenessZdual_coef_finitenessr*r*r+rs"          ($        zBaseLibSVM.fitcCst|ddjtjddS)zxValidation of y and class_weight. Default implementation for SVR and one-class; overridden in BaseSVC. TwarnF)r)rZastyper$rm)rPrr*r*r+rn'szBaseLibSVM._validate_targetscCs.|jdvsJ|jdkr*td|jtdS)NrrrznSolver terminated early (max_iter=%i). Consider pre-processing your data with StandardScaler or MinMaxScaler.) fit_status_warningsrrKr rPr*r*r+_warn_from_fit_status.s z BaseLibSVM._warn_from_fit_statuscCst|jr6||_||}|jd|jdkr6tdt|jtj ||||t |dt d||j |j|j|j|j|j|j|j|j|j|j|d\ |_|_|_|_|_|_|_|_|_ |!dS)Nrrz(X.shape[0] should be equal to X.shape[1] class_weight_)svm_typerrJr?rDrErHr@rGrCrIrBrArFrKrf)"rkr?_BaseLibSVM__Xfit_compute_kernelr#rOlibsvmset_verbosity_wrapr;rgetattrr$emptyrDrErHr@rGrCrIrBrrrFrKsupport_support_vectors_ _n_supportrr_probA_probBrrr)rPrrrrr?rfr*r*r+ry8sJ   zBaseLibSVM._dense_fitc CsPtj|jtjdd|_||j|}t|j t |j d|j|j |j ||||j|j|j|j|jt|dtd||j|j|jt|jt|j|j|\ |_|_}|_|_|_ |_!|_"|_#|$t%|drt&|j'd} nd} |jj d} t(t)| | } | st*+g|_,n2t)d| j-d| j-| } t*+|| | f| | f|_,dS)NrDr]r^rrrr).r$rorrm sort_indices_sparse_kernelsrp libsvm_sparserr;Zlibsvm_sparse_trainr#indicesindptrr@rrrBrCrDrrrErIrFintrGrHrKrrrrrrrrrr~rrZtileZarangerh csr_matrixrsize) rPrrrrr?rf kernel_typeZdual_coef_datar(Zn_SVZdual_coef_indicesZdual_coef_indptrr*r*r+rxgsf    zBaseLibSVM._sparse_fitcCs$||}|jr|jn|j}||S)aPerform regression on samples in X. For an one-class model, +1 (inlier) or -1 (outlier) is returned. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) For kernel="precomputed", the expected shape of X is (n_samples_test, n_samples_train). Returns ------- y_pred : ndarray of shape (n_samples,) The predicted values. )_validate_for_predictrl_sparse_predict_dense_predict)rPrpredictr*r*r+rs zBaseLibSVM.predictcCs||}|jdkr"t|ddd}|j}t|jrfd}|jd|jdkrftd|jd|jdft |j }t j ||j |j|j|j|j|j|j|||j|j|j|jdS) NrrDF)r^r`r0rMX.shape[1] = %d should be equal to %d, the number of samples at training time)rr?r@rBrArI)rndimr r?rkr#rrOrNrprMrrrrrrrrrr@rBrrrI)rPrr?rr*r*r+rs:    zBaseLibSVM._dense_predictcCs|j}t|rd}|j|}d}t|j|j|j|j j|j j|j j|j j|j t |j ||j|j|j|j|t|dtd|j|j|j|j|j|j|jS)Nr0r7rr)r?rkrrprZlibsvm_sparse_predictrrrrrrrNrMr@rrrBrCrr$rrErFrGrHrrr)rPrr?rrDr*r*r+rs<  zBaseLibSVM._sparse_predictcCs@t|jr<|||j}t|r*|}tj|tjdd}|S)z0Return the data transformed by a callable kernelrDr) rkr?rrhriZtoarrayr$rormrPrr?r*r*r+rs   zBaseLibSVM._compute_kernelcCsV||}||}|jr&||}n ||}|jdvrRt|jdkrR| S|S)afEvaluates the decision function for the samples in X. Parameters ---------- X : array-like of shape (n_samples, n_features) Returns ------- X : array-like of shape (n_samples, n_class * (n_class-1) / 2) Returns the decision function of the sample for each class in the model. rgr) rrrl_sparse_decision_function_dense_decision_functionrMrrravel)rPrZdec_funcr*r*r+_decision_function s     zBaseLibSVM._decision_functioncCsht|tjddd}|j}t|r$d}tj||j|j|j |j |j |j |j t|j||j|j|j|jdS)NrDF)r]r^r`r0rr?r@rIrBrA)r r$rmr?rkrdecision_functionrrrrrrrrNrprMr@rIrBrrrr*r*r+r's( z#BaseLibSVM._dense_decision_functioncCstj|jtjdd|_|j}t|dr*d}|j|}t |j|j |j |j j|j j |j j |j j|jt|j||j|j|j|j|jt|dtd|j|j|j|j|j|j|jS)NrDr__call__r0rr)r$rorrmr?r~rrprZlibsvm_sparse_decision_functionrrrrrrNrMr@rrrBrCrDrrrErFrGrHrrrrPrr?rr*r*r+r?s<   z$BaseLibSVM._sparse_decision_functionc Cst|t|js*t||dtjdddd}|jrDt|sDt |}|jrR| t|r~|js~t|js~t dt |j |jdkr|jd|jdkrt d |jd|jdf|j}|js|jdkr|j|jdkrt d |jj d |S) Nr\rDF)r_r]r^r`resetz3cannot use sparse input in %r trained on dense datar0rrrzThe internal representation of z was altered)rrkr?rr$rmrlrhrirrrOtype__name__r#rrr n_support_sumr[)rPrsvr*r*r+rbsD    $z BaseLibSVM._validate_for_predictcCs<|jdkrtd|}t|r0d|jj_nd|j_|S)zWeights assigned to the features when `kernel="linear"`. Returns ------- ndarray of shape (n_features, n_classes) r.z2coef_ is only available when using a linear kernelF)r?AttributeError _get_coefrhrirflagsZ writeablerPr)r*r*r+coef_s   zBaseLibSVM.coef_cCst|j|jS)N)rrrrr*r*r+rszBaseLibSVM._get_coefcCsTz t|Wnty"tYn0t|j}|dvr>|jSt|jdgSdS)z)Number of support vectors for each class.rrN) rr rrNrprMrr$array)rPrr*r*r+rs    zBaseLibSVM.n_support_)N)#r __module__ __qualname____doc__rrkrrrdictrL__annotations__rrrQrTr rrnrryrxrrrrrrrrpropertyrrr __classcell__r*r*rZr+r-EsZ    '   /="$ #) r-) metaclasscseZdZUdZiejeddhgdgdZeed<dD]Z e e q:e fdd Z d d Z d d ZfddZddZeeddZeeddZddZddZddZeddZeddZfd d!ZZS)"BaseSVCz!ABC for LibSVM-based classifiers.ovrovor:)decision_function_shape break_tiesrL)rFrEcs:||_||_tj|||||||d|| | | | | |ddS)Nr7r>)rrrSrQ)rPr?r@rArBrCrDrErGrHrIrJr;rKrr=rrZr*r+rQs&zBaseSVC.__init__cCslt|dd}t|tj|dd\}}t|j||d|_t|dkrTtdt|||_ tj |tj ddS) NTr)Zreturn_inverseclassesrrz>The number of classes has to be greater than one; got %d classrDr) rrr$uniquerrJrrrOrrorm)rPrZy_clsr*r*r+rns  zBaseSVC._validate_targetscCs>||}|jdkr:t|jdkr:t|dk| t|jS|S)a4Evaluate the decision function for the samples in X. Parameters ---------- X : array-like of shape (n_samples, n_features) The input samples. Returns ------- X : ndarray of shape (n_samples, n_classes * (n_classes-1) / 2) Returns the decision function of the sample for each class in the model. If decision_function_shape='ovr', the shape is (n_samples, n_classes). Notes ----- If decision_function_shape='ovo', the function values are proportional to the distance of the samples X to the separating hyperplane. If the exact distances are required, divide the function values by the norm of the weight vector (``coef_``). See also `this question `_ for further details. If decision_function_shape='ovr', the decision function is a monotonic transformation of ovo decision function. rrr)rrrrr)rPrdecr*r*r+rs zBaseSVC.decision_functioncsxt||jr |jdkr td|jrT|jdkrTt|jdkrTtj||dd}n t |}|j tj |tj dS)aPerform classification on samples in X. For an one-class model, +1 or -1 is returned. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) or (n_samples_test, n_samples_train) For kernel="precomputed", the expected shape of X is (n_samples_test, n_samples_train). Returns ------- y_pred : ndarray of shape (n_samples,) Class labels for samples in X. rz>break_ties must be False when decision_function_shape is 'ovo'rrr)Zaxisra)rrrrOrrr$ZargmaxrrSrZtakeroZintp)rPrrrZr*r+rs  zBaseSVC.predictcCs$|jstd|jdvr tddS)Nz5predict_proba is not available when probability=Falsergz0predict_proba only implemented for SVC and NuSVCT)rHrrMrr*r*r+ _check_proba=s zBaseSVC._check_probacCsD||}|jjdks"|jjdkr*td|jr6|jn|j}||S)aCompute probabilities of possible outcomes for samples in X. The model needs to have probability information computed at training time: fit with attribute `probability` set to True. Parameters ---------- X : array-like of shape (n_samples, n_features) For kernel="precomputed", the expected shape of X is (n_samples_test, n_samples_train). Returns ------- T : ndarray of shape (n_samples, n_classes) Returns the probability of the sample for each class in the model. The columns correspond to the classes in sorted order, as they appear in the attribute :term:`classes_`. 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. rzApredict_proba is not available when fitted with probability=False)rprobA_rprobB_r rl_sparse_predict_proba_dense_predict_proba)rPrZ pred_probar*r*r+ predict_probaFs zBaseSVC.predict_probacCst||S)aCompute 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 of shape (n_samples, n_features) or (n_samples_test, n_samples_train) For kernel="precomputed", the expected shape of X is (n_samples_test, n_samples_train). Returns ------- T : ndarray of shape (n_samples, n_classes) Returns the log-probabilities of the sample for each class in the model. The columns correspond to the classes in sorted order, as they appear in the attribute :term:`classes_`. 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. )r$logr)rPrr*r*r+predict_log_probakszBaseSVC.predict_log_probacCsh||}|j}t|rd}t|j}tj||j|j |j |j |j |j |j|||j|j|j|jd}|S)Nr0r)rr?rkrNrprMrrrrrrrrrr@rIrBrr)rPrr?rZpprobr*r*r+rs,  zBaseSVC._dense_predict_probacCstj|jtjdd|_|j}t|r(d}|j|}t |j|j |j |j j|j j |j j |j j|jt|j||j|j|j|j|jt|dtd|j|j|j|j|j|j|jS)NrDrr0rr)r$rorrmr?rkrrprZlibsvm_sparse_predict_probarrrrrrNrMr@rrrBrCrDrrrErFrGrHrrrrr*r*r+rs<  zBaseSVC._sparse_predict_probacCs^|jjddkr t|j|j}n:t|j|j|j}t|drPt| }n t |}|S)Nrr) rr#rrr,rrhriZvstackZtocsrr$rr*r*r+rs  zBaseSVC._get_coefcCs|jSzParameter learned in Platt scaling when `probability=True`. Returns ------- ndarray of shape (n_classes * (n_classes - 1) / 2) )rrr*r*r+rszBaseSVC.probA_cCs|jSr)rrr*r*r+rszBaseSVC.probB_cst}|jdk|j_|SrR)rSrTr?rUrWrXrZr*r+rTs zBaseSVC.__sklearn_tags__)rrrrr-rLrrrZ unused_parampoprrQrnrrrrrrrrrrrrrTrr*r*rZr+rs6   ' %  $ #  rc Csddiddddddd iidd id d ddddd iiddddidd}|dkr\||S|dkrptd|||d}|durd|}nJ||d}|durd||f}n(||d}|durd|||f}n|Std||||fdS)aFind the liblinear magic number for the solver. This number depends on the values of the following attributes: - multi_class - penalty - loss - dual The same number is also internally used by LibLinear to determine which solver to use. Fr)FT)l1l2rTrr )logistic_regressionZhingeZ squared_hingeepsilon_insensitivesquared_epsilon_insensitivecrammer_singerrrz<`multi_class` must be one of `ovr`, `crammer_singer`, got %rNzloss='%s' is not supportedz>The combination of penalty='%s' and loss='%s' is not supportedzLThe combination of penalty='%s' and loss='%s' are not supported when dual=%szJUnsupported set of arguments: %s, Parameters: penalty=%r, loss=%r, dual=%r)rOget) multi_classpenaltylossdualZ_solver_type_dictZ _solver_penZ error_stringZ _solver_dualZ solver_numr*r*r+_get_liblinear_solver_typesD         rrr皙?cCs| dvrJt}||}|j}t|dkr:td|dt|||d}ntjdtjd}|}t |t | }|rt ddd d }|r|dkrtd |n|}t |t |t |t|rt|tj|tjd}tj|d d }t||tjd}t| || |}t ||t||| |||| |tdj|| \}}t|}|| krhtdt|r|ddddf}||dddf}n|}d}|||fS)aUsed by Logistic Regression (and CV) and LinearSVC/LinearSVR. Preprocessing is done in this function before supplying it to liblinear. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) Training vector, where `n_samples` is the number of samples and `n_features` is the number of features. y : array-like of shape (n_samples,) Target vector relative to X C : float Inverse of cross-validation parameter. The lower the C, the higher the penalization. fit_intercept : bool Whether or not to fit an intercept. If set to True, the feature vector is extended to include an intercept term: ``[x_1, ..., x_n, 1]``, where 1 corresponds to the intercept. If set to False, no intercept will be used in calculations (i.e. data is expected to be already centered). intercept_scaling : float Liblinear internally penalizes the intercept, treating it like any other term in the feature vector. To reduce the impact of the regularization on the intercept, the `intercept_scaling` parameter can be set to a value greater than 1; the higher the value of `intercept_scaling`, the lower the impact of regularization on it. Then, the weights become `[w_x_1, ..., w_x_n, w_intercept*intercept_scaling]`, where `w_x_1, ..., w_x_n` represent the feature weights and the intercept weight is scaled by `intercept_scaling`. This scaling allows the intercept term to have a different regularization behavior compared to the other features. class_weight : dict or 'balanced', default=None Weights associated with classes in the form ``{class_label: weight}``. If not given, all classes are supposed to have weight one. For multi-output problems, a list of dicts can be provided in the same order as the columns of y. The "balanced" mode uses the values of y to automatically adjust weights inversely proportional to class frequencies in the input data as ``n_samples / (n_classes * np.bincount(y))`` penalty : {'l1', 'l2'} The norm of the penalty used in regularization. dual : bool Dual or primal formulation, verbose : int Set verbose to any positive number for verbosity. max_iter : int Number of iterations. tol : float Stopping condition. random_state : int, RandomState instance or None, default=None Controls the pseudo random number generation for shuffling the data. Pass an int for reproducible output across multiple function calls. See :term:`Glossary `. multi_class : {'ovr', 'crammer_singer'}, default='ovr' `ovr` trains n_classes one-vs-rest classifiers, while `crammer_singer` optimizes a joint objective over all classes. While `crammer_singer` is interesting from an theoretical perspective as it is consistent it is seldom used in practice and rarely leads to better accuracy and is more expensive to compute. If `crammer_singer` is chosen, the options loss, penalty and dual will be ignored. loss : {'logistic_regression', 'hinge', 'squared_hinge', 'epsilon_insensitive', 'squared_epsilon_insensitive}, default='logistic_regression' The loss function used to fit the model. epsilon : float, default=0.1 Epsilon parameter in the epsilon-insensitive loss function. Note that the value of this parameter depends on the scale of the target variable y. If unsure, set epsilon=0. sample_weight : array-like of shape (n_samples,), default=None Weights assigned to each sample. Returns ------- coef_ : ndarray of shape (n_features, n_features + 1) The coefficient vector got by minimizing the objective function. intercept_ : float The intercept term added to the vector. n_iter_ : array of int Number of iterations run across for each class. )rrrzeThis solver needs samples of at least 2 classes in the data, but the data contains only one class: %rrrraz [LibLinear]rbrcgzqIntercept scaling is %r but needs to be greater than 0. To disable fitting an intercept, set fit_intercept=False.W) requirementsrez@Liblinear failed to converge, increase the number of iterations.Nr<r7)r Z fit_transformrrrOrr$rrm liblinearrrrzrrrhrirrorrequirerrZ train_wrapr{r|r}rrr )rrrDZ fit_interceptZintercept_scalingrJrrr;rKrCr=rrrFrencZy_indrrrZbiasrZ raw_coef_rZ n_iter_maxrrr*r*r+_fit_liblinear*szt         r)NrrrN)8rabcrrnumbersrrnumpyr$Z scipy.sparserWrhbaserrr exceptionsr r Z preprocessingr utilsr rrrZutils._param_validationrrZ utils.extmathrZutils.metaestimatorsrZutils.multiclassrrZutils.validationrrrrrrrbrrrrr rrNr,r-rrrr*r*r*r+sB         u=E