Machine Learning with scikit-learn: The Complete Guide

Core Concepts:

• Estimator API: Consistent fit()/predict() pattern
• Pipelines: Chaining preprocessing and modeling
• Transformers: fit_transform() for feature processing
• Model Persistence: pickle or joblib

Basic Workflow:

from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

# Load data
X, y = load_data()

# Split data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

# Train model
model = RandomForestClassifier(n_estimators=100)
model.fit(X_train, y_train)

# Evaluate
preds = model.predict(X_test)
print(f"Accuracy: {accuracy_score(y_test, preds):.2f}")
        

Common Transformers:

Pipeline Example:

from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import StandardScaler, OneHotEncoder

# Define preprocessing
numeric_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='median')),
    ('scaler', StandardScaler())])

categorical_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='constant', fill_value='missing')),
    ('onehot', OneHotEncoder(handle_unknown='ignore'))])

preprocessor = ColumnTransformer(
    transformers=[
        ('num', numeric_transformer, numeric_features),
        ('cat', categorical_transformer, categorical_features)])

# Full pipeline
pipeline = Pipeline(steps=[
    ('preprocessor', preprocessor),
    ('classifier', RandomForestClassifier())])
        

Tuning Methods:

Implementation:

from sklearn.model_selection import GridSearchCV

# Define parameter grid
param_grid = {
    'preprocessor__num__imputer__strategy': ['mean', 'median'],
    'classifier__n_estimators': [50, 100, 200],
    'classifier__max_depth': [None, 5, 10]
}

# Search
search = GridSearchCV(pipeline, param_grid, cv=5)
search.fit(X_train, y_train)

print(f"Best params: {search.best_params_}")
print(f"Best score: {search.best_score_:.3f}")
        

Powerful Features:

• Custom Transformers: Create your own with BaseEstimator
• Feature Unions: Combine multiple feature extraction methods
• Metric Scoring: Define custom evaluation metrics
• Out-of-Core: Partial_fit for large datasets

Custom Transformer Example:

from sklearn.base import BaseEstimator, TransformerMixin

class TextLengthTransformer(BaseEstimator, TransformerMixin):
    def fit(self, X, y=None):
        return self
    def transform(self, X):
        return [[len(text)] for text in X]
        
# Use in pipeline
pipeline = Pipeline([
    ('text_len', TextLengthTransformer()),
    ('clf', RandomForestClassifier())
])