sklearn 线性回归_sklearn-部分API

（本文没有任何参考价值，纯属个人笔记）

sklearn.feature_extraction

• DictVectorizer : 特征转化，将特征名和特征值组合形成新的特征名
• image
• img_to_graph
• grid_to_graph
• text : { CountVectorizer=>统计每篇文章词频次数， TfidfVectorizer => 计算词的重要性}
• FeatureHasher

sklearn.preprocessing

• 'Binarizer',
• 'FunctionTransformer',
• 'Imputer', ： 缺失值补充 ，im = Imputer(missing_values='NaN', strategy='mean', axis=0)

• 'KernelCenterer',
• 'LabelBinarizer',
• 'LabelEncoder',
• 'MultiLabelBinarizer',
• 'MinMaxScaler', : 归一化，把值限定在一定范围里 比如 MinMaxScaler(feature_range=(0, 1))，最小的值为 0 最大的值为1，其余的按比例缩减
• 'MaxAbsScaler',
• 'QuantileTransformer',
• 'Normalizer',
• 'OneHotEncoder',
• 'RobustScaler',
• 'StandardScaler', ： 标准化，数据中有如果出现异常点，比如十分夸张的极值，势必会对预测结果准确率造成较大影响，通过标准化计算标准差，去除异常值的点。

• 'add_dummy_feature',
• 'PolynomialFeatures',
• 'binarize',
• 'normalize',
• 'scale',
• 'robust_scale',
• 'maxabs_scale',
• 'minmax_scale',
• 'label_binarize',
• 'quantile_transform',

sklearn.feature_selection

• GenericUnivariateSelect
• RFE
• RFECV
• SelectFdr
• SelectFpr
• SelectFwe
• SelectKBest
• SelectFromModel
• SelectPercentile
• VarianceThreshold ： 过滤式，过滤掉方差比较小的特征，方差比较小说明，该特征变化幅度比较小。
• chi2
• f_classif
• f_oneway
• f_regression
• mutual_info_classif
• mutual_info_regression

sklearn.decomposition

• DictionaryLearning
• FastICA
• IncrementalPCA
• KernelPCA
• MiniBatchDictionaryLearning
• MiniBatchSparsePCA
• NMF
• PCA : 主成分分析，用于简化数据集。

• RandomizedPCA
• SparseCoder
• SparsePCA
• dict_learning
• dict_learning_online
• fastica
• non_negative_factorization
• randomized_svd
• sparse_encode
• FactorAnalysis
• TruncatedSVD
• LatentDirichletAllocation

sklearn.neighbors

• BallTree
• DistanceMetric
• KDTree
• KNeighborsClassifier
• KNeighborsRegressor ： K临近值算法实体
• NearestCentroid
• NearestNeighbors
• RadiusNeighborsClassifier
• RadiusNeighborsRegressor
• kneighbors_graph
• radius_neighbors_graph
• KernelDensity
• LSHForest
• LocalOutlierFactor

sklearn.model_selection

• BaseCrossValidator
• GridSearchCV : 交叉验证

• TimeSeriesSplit
• KFold
• GroupKFold
• GroupShuffleSplit
• LeaveOneGroupOut
• LeaveOneOut
• LeavePGroupsOut
• LeavePOut
• RepeatedKFold
• RepeatedStratifiedKFold
• ParameterGrid
• ParameterSampler
• PredefinedSplit
• RandomizedSearchCV
• ShuffleSplit
• StratifiedKFold
• StratifiedShuffleSplit
• check_cv
• cross_val_predict
• cross_val_score
• cross_validate
• fit_grid_point
• learning_curve
• permutation_test_score
• train_test_split : 训练集和测试集拆分
• validation_curve

sklearn.tree

• DecisionTreeClassifier : 决策树

• DecisionTreeRegressor
• ExtraTreeClassifier
• ExtraTreeRegressor
• export_graphviz : 决策树的可视化

sklearn.ensemble

• BaseEnsemble
• RandomForestClassifier : 随机森林算法，使用交叉验证，网格搜索确定最优参数组合。

• RandomForestRegressor
• RandomTreesEmbedding
• ExtraTreesClassifier
• ExtraTreesRegressor
• BaggingClassifier
• BaggingRegressor
• IsolationForest
• GradientBoostingClassifier
• GradientBoostingRegressor
• AdaBoostClassifier
• AdaBoostRegressor
• VotingClassifier
• bagging
• forest
• gradient_boosting
• partial_dependence
• weight_boosting

sklearn.naive_bayes

• BernoulliNB
• GaussianNB
• MultinomialNB ： 朴素贝叶斯分类算法

sklearn.linear_model

• ARDRegression
• BayesianRidge
• ElasticNet
• ElasticNetCV
• Hinge
• Huber
• HuberRegressor
• Lars
• LarsCV
• Lasso
• LassoCV
• LassoLars
• LassoLarsCV
• LassoLarsIC
• LinearRegression :线性回归， 采用正规方程求最优解https://zhuanlan.zhihu.com/p/87866712
• Log
• LogisticRegression
• LogisticRegressionCV
• ModifiedHuber
• MultiTaskElasticNet
• MultiTaskElasticNetCV
• MultiTaskLasso
• MultiTaskLassoCV
• OrthogonalMatchingPursuit
• OrthogonalMatchingPursuitCV
• PassiveAggressiveClassifier
• PassiveAggressiveRegressor
• Perceptron
• RandomizedLasso
• RandomizedLogisticRegression
• Ridge ： 岭回归（使用正则化的线性回归）https://zhuanlan.zhihu.com/p/87866712
• RidgeCV
• RidgeClassifier
• RidgeClassifierCV
• SGDClassifier
• SGDRegressor ：线性回归， 采用最小二乘法计算最优值，https://zhuanlan.zhihu.com/p/87866712
• SquaredLoss
• TheilSenRegressor
• enet_path
• lars_path
• lasso_path
• lasso_stability_path
• logistic_regression_path
• orthogonal_mp
• orthogonal_mp_gram
• ridge_regression
• RANSACRegressor

sklearn.metrics

• accuracy_score
• adjusted_mutual_info_score
• adjusted_rand_score
• auc
• average_precision_score
• calinski_harabaz_score
• classification_report
• cluster
• cohen_kappa_score
• completeness_score
• confusion_matrix
• consensus_score
• coverage_error
• euclidean_distances
• explained_variance_score
• f1_score
• fbeta_score
• fowlkes_mallows_score
• get_scorer
• hamming_loss
• hinge_loss
• homogeneity_completeness_v_measure
• homogeneity_score
• jaccard_similarity_score
• label_ranking_average_precision_score
• label_ranking_loss
• log_loss
• make_scorer
• matthews_corrcoef
• mean_absolute_error
• mean_squared_error ： 计算均方误差
• mean_squared_log_error
• median_absolute_error
• mutual_info_score
• normalized_mutual_info_score
• pairwise_distances
• pairwise_distances_argmin
• pairwise_distances_argmin_min
• pairwise_kernels
• precision_recall_curve
• precision_recall_fscore_support
• precision_score
• r2_score
• recall_score
• roc_auc_score
• roc_curve
• SCORERS
• silhouette_samples
• silhouette_score
• v_measure_score
• zero_one_loss
• brier_score_loss
• dcg_score
• ndcg_score