Scikit-Learn

• Base class
  • estimator
  • classifier
  • cluster
  • regressor
  • transformer
• Datasets
  • datasets.load_svmlight_files(files, n_features, dtype)
• Cluster
  • cluster.KMeans(n_cluster, max_iter, n_init, init:{kmeans++,random}).fit(X).predict(X)
  • cluster.DBSCAN(eps, min_scample, m etric, algorithm: {auto, ball_tree, kd_tree}).fit(X).predict(X)
• Matrix Decomposition
  • decomposition.NMF(n_components, init_method, solver: {'pg', 'cd'}, tolerance, max_iter, alpha, l1_ratio).fit(X).transform(X)
• Ensemble
  • ensemble.GradientBoostingClassifiler(loss: {logloss, expo}, learning_rate, n_trees, max_depth, criterion: {mse, mae}, min_split_samples, min_leaf_samples, min_leaf_weight, subsample, max_features, max_leaf_nodes).fit(X,y).predict(X)
  • ensemble.GradientBoostingRegressor()
  • ensemble.RandomForestClassifier(n_trees, criterion: {gini, entropy}, max_features, max_depth, min_split_samples, min_leaf_samples, max_leaf_nodes).fit(X,y).predict(X)
  • ensemble.RandomForestRegressor()
• Generalized Linear Model
  • linear_model.LinearRegression(fit_intercept, normalize).fit(X,y).predict(X,y)
  • linear_model.LogisticRegression(penalty: {l1, l2}, fit_intercept, max_iter, solver: {newton, lbfgs, liblinear, sag}, tolerance).fit(X,y),predict(X)
  • linear_model.lasso
  • linear_model.SGDClassifiler(loss: {hinge, log, squared_loss}, penalty, alpha, l1_ratio, fit_intercept, max_iter, shuffle, learning_rate: {constant, optimal, invscaling}, eta0, power_t).fit(X,y),predict(X)
• Metrics
  • metrics.accuracy_score(y_true,y_pred)
  • metrics.auc
  • metrics.f1_score
  • metrics.hinge_loss
  • metrics.log_loss
  • metrics.precision_recall_curve
  • metrics.roc
  • metrics.mean_absolute_error
  • metrics.mean_squared_error
• Pipeline
  • pipeline.Pipeline(steps)
• Preprocessing
  • preprocessing.MaxAbsScaler
  • preprocessing.Normalizer
  • preprocessing.OneHotEncoder
• Support Vector Machine

Keras

• Input
• Dense
• Model(input, output).compile(optimizer,loss,metric).fit().evaluate().predict()
• Optimizer
• Loss
• Metric

PyTorch

• Tensor
• Storage
• Optim