0Ph-dZddlmZmZddlZddlmZddl m Z ddl m Z m Z mZmZddlmZmZddlmZdd lmZmZdd lmZmZmZdd lmZdd lmZm Z d gZ!Gdd e ee Z"dS)zFTruncated SVD for sparse matrices, aka latent semantic analysis (LSA).)IntegralRealN)svds) BaseEstimatorClassNamePrefixFeaturesOutMixinTransformerMixin _fit_context) check_arraycheck_random_state)_init_arpack_v0)Interval StrOptions)randomized_svdsafe_sparse_dotsvd_flip)mean_variance_axis)check_is_fitted validate_data TruncatedSVDc TeZdZUdZeedddgeddhgeedddgeedddgehd gd geedddgd Ze e d < dddddddddZ ddZ e dddZdZdZfdZedZxZS)raDimensionality reduction using truncated SVD (aka LSA). This transformer performs linear dimensionality reduction by means of truncated singular value decomposition (SVD). Contrary to PCA, this estimator does not center the data before computing the singular value decomposition. This means it can work with sparse matrices efficiently. In particular, truncated SVD works on term count/tf-idf matrices as returned by the vectorizers in :mod:`sklearn.feature_extraction.text`. In that context, it is known as latent semantic analysis (LSA). This estimator supports two algorithms: a fast randomized SVD solver, and a "naive" algorithm that uses ARPACK as an eigensolver on `X * X.T` or `X.T * X`, whichever is more efficient. Read more in the :ref:`User Guide `. Parameters ---------- n_components : int, default=2 Desired dimensionality of output data. If algorithm='arpack', must be strictly less than the number of features. If algorithm='randomized', must be less than or equal to the number of features. The default value is useful for visualisation. For LSA, a value of 100 is recommended. algorithm : {'arpack', 'randomized'}, default='randomized' SVD solver to use. Either "arpack" for the ARPACK wrapper in SciPy (scipy.sparse.linalg.svds), or "randomized" for the randomized algorithm due to Halko (2009). n_iter : int, default=5 Number of iterations for randomized SVD solver. Not used by ARPACK. The default is larger than the default in :func:`~sklearn.utils.extmath.randomized_svd` to handle sparse matrices that may have large slowly decaying spectrum. n_oversamples : int, default=10 Number of oversamples for randomized SVD solver. Not used by ARPACK. See :func:`~sklearn.utils.extmath.randomized_svd` for a complete description. .. versionadded:: 1.1 power_iteration_normalizer : {'auto', 'QR', 'LU', 'none'}, default='auto' Power iteration normalizer for randomized SVD solver. Not used by ARPACK. See :func:`~sklearn.utils.extmath.randomized_svd` for more details. .. versionadded:: 1.1 random_state : int, RandomState instance or None, default=None Used during randomized svd. Pass an int for reproducible results across multiple function calls. See :term:`Glossary `. tol : float, default=0.0 Tolerance for ARPACK. 0 means machine precision. Ignored by randomized SVD solver. Attributes ---------- components_ : ndarray of shape (n_components, n_features) The right singular vectors of the input data. explained_variance_ : ndarray of shape (n_components,) The variance of the training samples transformed by a projection to each component. explained_variance_ratio_ : ndarray of shape (n_components,) Percentage of variance explained by each of the selected components. singular_values_ : ndarray of shape (n_components,) The singular values corresponding to each of the selected components. The singular values are equal to the 2-norms of the ``n_components`` variables in the lower-dimensional space. n_features_in_ : int Number of features seen during :term:`fit`. .. versionadded:: 0.24 feature_names_in_ : ndarray of shape (`n_features_in_`,) Names of features seen during :term:`fit`. Defined only when `X` has feature names that are all strings. .. versionadded:: 1.0 See Also -------- DictionaryLearning : Find a dictionary that sparsely encodes data. FactorAnalysis : A simple linear generative model with Gaussian latent variables. IncrementalPCA : Incremental principal components analysis. KernelPCA : Kernel Principal component analysis. NMF : Non-Negative Matrix Factorization. PCA : Principal component analysis. Notes ----- SVD suffers from a problem called "sign indeterminacy", which means the sign of the ``components_`` and the output from transform depend on the algorithm and random state. To work around this, fit instances of this class to data once, then keep the instance around to do transformations. References ---------- :arxiv:`Halko, et al. (2009). "Finding structure with randomness: Stochastic algorithms for constructing approximate matrix decompositions" <0909.4061>` Examples -------- >>> from sklearn.decomposition import TruncatedSVD >>> from scipy.sparse import csr_matrix >>> import numpy as np >>> np.random.seed(0) >>> X_dense = np.random.rand(100, 100) >>> X_dense[:, 2 * np.arange(50)] = 0 >>> X = csr_matrix(X_dense) >>> svd = TruncatedSVD(n_components=5, n_iter=7, random_state=42) >>> svd.fit(X) TruncatedSVD(n_components=5, n_iter=7, random_state=42) >>> print(svd.explained_variance_ratio_) [0.0157... 0.0512... 0.0499... 0.0479... 0.0453...] >>> print(svd.explained_variance_ratio_.sum()) 0.2102... >>> print(svd.singular_values_) [35.2410... 4.5981... 4.5420... 4.4486... 4.3288...] Nleft)closedarpack randomizedr>LUORautonone random_state) n_components algorithmn_iter n_oversamplespower_iteration_normalizerr!tol_parameter_constraintsr rg)r#r$r%r&r!r'ch||_||_||_||_||_||_||_dSN)r#r"r$r%r&r!r')selfr"r#r$r%r&r!r's d/var/www/html/test/jupyter/venv/lib/python3.11/site-packages/sklearn/decomposition/_truncated_svd.py__init__zTruncatedSVD.__init__s=#( **D'(c0|||S)aoFit model on training data X. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) Training data. y : Ignored Not used, present here for API consistency by convention. Returns ------- self : object Returns the transformer object. ) fit_transform)r-Xys r.fitzTruncatedSVD.fits 1 r0T)prefer_skip_nested_validationc ,t||ddgd}t|j}|jdkrzt t |j|}t||j|j |\}}}|ddd}t|dddddf|dddd \}}n|jd kr|j|jd kr&td |jd|jd dt||j|j |j|j|d \}}}t||d \}}||_|jd ks|jdkr&|j dkrt#||jj}n||z}t'j|dx|_} t-j|r)t1|d\} } | } n(t'j|d} | | z |_||_|S)aFit model to X and perform dimensionality reduction on X. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) Training data. y : Ignored Not used, present here for API consistency by convention. Returns ------- X_new : ndarray of shape (n_samples, n_components) Reduced version of X. This will always be a dense array. csrcscr) accept_sparseensure_min_featuresr)kr'v0NF)u_based_decisionrrz n_components(z) must be <= n_features(z).)r$r%r&r! flip_signr)axis)rr r!r#r minshaperr"r'r ValueErrorrr$r%r& components_rTnpvarexplained_variance_spissparsersumexplained_variance_ratio_singular_values_) r-r3r4r!r=USigmaVT X_transformedexp_var_full_vars r.r2zTruncatedSVD.fit_transformsL" $%UV W W W)$*;<< >X % % QW|<\ ) ) Nh & &48a<<+At/?/ABBMMIM.0VM-J-J-JJ 7 ;q>> /,QQ777KAx||~~HHvaa(((,,..H)08);& %r0c~t|t||ddgd}t||jjS)aVPerform dimensionality reduction on X. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) New data. Returns ------- X_new : ndarray of shape (n_samples, n_components) Reduced version of X. This will always be a dense array. r8r9F)r:reset)rrrrErFr-r3s r. transformzTruncatedSVD.transformsB  $%u M M Mq$"2"4555r0cTt|}tj||jS)a{Transform X back to its original space. Returns an array X_original whose transform would be X. Parameters ---------- X : array-like of shape (n_samples, n_components) New data. Returns ------- X_original : ndarray of shape (n_samples, n_features) Note that this is always a dense array. )r rGdotrErXs r.inverse_transformzTruncatedSVD.inverse_transform's$ NNva)***r0c|t}d|j_ddg|j_|S)NTfloat64float32)super__sklearn_tags__ input_tagssparsetransformer_tagspreserves_dtype)r-tags __class__s r.razTruncatedSVD.__sklearn_tags__9s7ww''))!%1:I0F- r0c&|jjdS)z&Number of transformed output features.r)rErC)r-s r._n_features_outzTruncatedSVD._n_features_out?s%a((r0)rr,)__name__ __module__ __qualname____doc__rrrrr(dict__annotations__r/r5r r2rYr\rapropertyri __classcell__)rgs@r.rrsBBJ"(AtFCCCD j(L!9::;8Haf===>"(8QVDDDE'1z2N2N2N'O'O&P'(q$v6667$$D#) &&\555BBB65BH666"+++$ ))X)))))r0)#rmnumbersrrnumpyrG scipy.sparsercrJscipy.sparse.linalgrbaserrr r utilsr r utils._arpackr utils._param_validationrr utils.extmathrrrutils.sparsefuncsrutils.validationrr__all__rr0r.rsiLL #"""""""$$$$$$ 43333333++++++::::::::EEEEEEEEEE222222========  f)f)f)f)f)24Dmf)f)f)f)f)r0