Scikit learn

A concise and powerful cheatsheet for

Setup & Basics

TaskCodeDescription
Import sklearnimport sklearnImport library
Check versionsklearn.versionDisplay version
Import dataset utilitiesfrom sklearn import datasetsBuilt-in datasets
Import model selectionfrom sklearn.model_selection import train_test_splitSplit dataset
Import metricsfrom sklearn import metricsEvaluation metrics
Import preprocessingfrom sklearn import preprocessingData preprocessing

Loading Built-in Datasets

TaskCodeDescription
Load Iris datasetfrom sklearn.datasets import load_iris <br> iris = load_iris()Classic dataset
Load Digits datasetfrom sklearn.datasets import load_digitsHandwritten digits
Load Wine datasetfrom sklearn.datasets import load_wineWine classification
Load Breast Cancer datasetfrom sklearn.datasets import load_breast_cancerCancer detection
Split dataX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)Train/test split

Preprocessing

TaskCodeDescription
Standard scalingfrom sklearn.preprocessing import StandardScaler <br> scaler = StandardScaler()<br>X_scaled = scaler.fit_transform(X)Normalize features (mean=0, std=1)
Min-Max scalingfrom sklearn.preprocessing import MinMaxScalerScale values between 0–1
Label encodingfrom sklearn.preprocessing import LabelEncoder<br>le = LabelEncoder()Convert labels to integers
One-hot encodingfrom sklearn.preprocessing import OneHotEncoderConvert categorical to binary columns
Binarizationfrom sklearn.preprocessing import Binarizer(threshold=0.5)Convert values above threshold to 1
Polynomial featuresfrom sklearn.preprocessing import PolynomialFeatures(degree=2)Generate interaction terms

Feature Selection

TaskCodeDescription
Variance thresholdfrom sklearn.feature_selection import VarianceThresholdRemove low-variance features
Select K bestfrom sklearn.feature_selection import SelectKBest, f_classifKeep top K features
Recursive feature eliminationfrom sklearn.feature_selection import RFERecursive feature selection
Feature importancesmodel.feature_importances_Get importance after fitting

Supervised Learning Models

AlgorithmImportDescription
Linear Regressionfrom sklearn.linear_model import LinearRegressionPredict continuous values
Logistic Regressionfrom sklearn.linear_model import LogisticRegressionBinary classification
Decision Treefrom sklearn.tree import DecisionTreeClassifierTree-based model
Random Forestfrom sklearn.ensemble import RandomForestClassifierEnsemble of trees
Support Vector Machinefrom sklearn.svm import SVCMax-margin classifier
Naive Bayesfrom sklearn.naive_bayes import GaussianNBProbabilistic model
K-Nearest Neighborsfrom sklearn.neighbors import KNeighborsClassifierDistance-based classifier
Gradient Boostingfrom sklearn.ensemble import GradientBoostingClassifierBoosted trees
AdaBoostfrom sklearn.ensemble import AdaBoostClassifierAdaptive boosting
Linear SVMfrom sklearn.svm import LinearSVCFast linear classifier

Unsupervised Learning Models

AlgorithmImportDescription
K-Means Clusteringfrom sklearn.cluster import KMeansPartition data into clusters
Hierarchical Clusteringfrom sklearn.cluster import AgglomerativeClusteringTree-based clustering
DBSCANfrom sklearn.cluster import DBSCANDensity-based clustering
Gaussian Mixturefrom sklearn.mixture import GaussianMixtureProbabilistic clustering
PCA (Dim. Reduction)from sklearn.decomposition import PCAReduce features
t-SNEfrom sklearn.manifold import TSNEVisualization for high-dim data

Model Training & Prediction

TaskCodeDescription
Train modelmodel.fit(X_train, y_train)Fit model to data
Make predictiony_pred = model.predict(X_test)Predict labels
Predict probabilitymodel.predict_proba(X_test)Class probabilities
Model coefficientsmodel.coef_Coefficients (linear models)
Model interceptmodel.intercept_Intercept term

Model Evaluation

MetricCodeDescription
Accuracymetrics.accuracy_score(y_test, y_pred)Correct predictions
Precisionmetrics.precision_score(y_test, y_pred)Positive predictive value
Recallmetrics.recall_score(y_test, y_pred)Sensitivity
F1 Scoremetrics.f1_score(y_test, y_pred)Harmonic mean
Confusion Matrixmetrics.confusion_matrix(y_test, y_pred)Actual vs predicted
Classification Reportprint(metrics.classification_report(y_test, y_pred))Summary of metrics
ROC Curvemetrics.roc_curve(y_test, y_score)True/false positive trade-off
AUC Scoremetrics.roc_auc_score(y_test, y_score)Area under ROC

Regression Metrics

MetricCodeDescription
Mean Absolute Errormetrics.mean_absolute_error(y_true, y_pred)Avg. absolute error
Mean Squared Errormetrics.mean_squared_error(y_true, y_pred)Avg. squared error
RMSEnp.sqrt(metrics.mean_squared_error(y_true, y_pred))Root MSE
R² Scoremetrics.r2_score(y_true, y_pred)Coefficient of determination

Model Selection & Validation

TaskCodeDescription
Train/Test splittrain_test_split(X, y, test_size=0.2)Random data split
K-Fold CVfrom sklearn.model_selection import KFoldK-fold cross-validation
Cross-val scorefrom sklearn.model_selection import cross_val_score<br>cross_val_score(model, X, y, cv=5)Evaluate model stability
Grid Searchfrom sklearn.model_selection import GridSearchCVHyperparameter tuning
Random Searchfrom sklearn.model_selection import RandomizedSearchCVFaster hyperparameter tuning

Pipeline

TaskCodeDescription
Import pipelinefrom sklearn.pipeline import PipelineChain steps together
Create pipelinepipe = Pipeline([('scaler', StandardScaler()), ('model', LogisticRegression())])Combine preprocessing + model
Fit pipelinepipe.fit(X_train, y_train)Train pipeline
Predict pipelinepipe.predict(X_test)Predict via pipeline
Evaluate pipelinepipe.score(X_test, y_test)Score via pipeline

Dimensionality Reduction

TaskCodeDescription
PCAfrom sklearn.decomposition import PCA<br>pca = PCA(n_components=2)Principal Component Analysis
Fit/TransformX_pca = pca.fit_transform(X)Apply PCA
Explained variancepca.explained_variance_ratio_Variance explained by each PC
LDAfrom sklearn.discriminant_analysis import LinearDiscriminantAnalysisLinear discriminant analysis

Visualization

TaskCodeDescription
Confusion matrix plotfrom sklearn.metrics import ConfusionMatrixDisplayVisual confusion matrix
ROC curvefrom sklearn.metrics import RocCurveDisplayPlot ROC curve
Precision-Recall curvefrom sklearn.metrics import PrecisionRecallDisplayPrecision vs recall
Plot decision boundaryplt.contourf()Visualize classification boundary

Common Workflows

StepCode Example
1️⃣ Load datasetX, y = datasets.load_iris(return_X_y=True)
2️⃣ Split dataX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
3️⃣ Choose modelmodel = RandomForestClassifier()
4️⃣ Train modelmodel.fit(X_train, y_train)
5️⃣ Predicty_pred = model.predict(X_test)
6️⃣ Evaluatemetrics.accuracy_score(y_test, y_pred)

Common Hyperparameters

ModelKey Parameters
LogisticRegressionpenalty, C, solver, max_iter
DecisionTreeClassifiercriterion, max_depth, min_samples_split
RandomForestClassifiern_estimators, max_depth, random_state
SVCkernel, C, gamma
KNeighborsClassifiern_neighbors, weights, metric
GradientBoostingClassifiern_estimators, learning_rate, max_depth