Scikit-learn is a Python package for training, testing, and evaluating classical machine-learning models. It turns prepared features and labels into fitted models, including decision trees and support vector machines. section 9.4 section 13.7
A typical workflow uses Scikit-learn to:
- Fit a decision tree to feature columns and labels. section 9.4
- Train a linear SVM and adjust its C parameter to balance margin width against classification mistakes. section 11.2
- Prepare numerical features, encode categorical features, select useful inputs, and compare hyperparameters with grid search and k-fold cross-validation. section 13.7
A small decision-tree experiment
Put two features, x_0 and x_1, in a DataFrame, store class labels in y, then fit DecisionTreeClassifier(). The resulting depth-two tree creates rectangular regions using thresholds such as x_0 = 5 and x_1 = 8 or 2.5. Try criterion='entropy' and compare the boundary. section 9.4
decision_tree = DecisionTreeClassifier() decision_tree.fit(features, labels)
Inspect the compact fit call that trains a decision tree from prepared features and labels. section 9.4
