Classical Methods

Overview

The classical methods module provides implementations of traditional machine learning algorithms for tabular data analysis. All methods inherit from the base classical_methods class and provide a unified interface for training, prediction, and evaluation.

Available Methods

Base Components:

• Base: Base class for all classical machine learning methods, providing common interface and utilities

Tree-Based Methods:

• Random Forest: Ensemble learning method using multiple decision trees

• XGBoost: Gradient boosting framework with optimized implementation

• LightGBM: Light gradient boosting machine with high efficiency

• CatBoost: Gradient boosting with categorical features support

Linear Methods:

• Logistic Regression: Linear model for classification tasks

• Linear Regression: Linear model for regression tasks

• Support Vector Machine: SVM classifier with kernel methods

Distance-Based Methods:

• K-Nearest Neighbors: Instance-based learning using nearest neighbors

• Nearest Centroid Method: Classification based on centroid distances

Probabilistic Methods:

• Naive Bayes: Probabilistic classifier based on Bayes theorem

Utility Methods:

• Dummy Classifier: Baseline classifier for comparison and testing

Classical Methods:

• Classical Methods
  • Overview
  • Available Methods
  • Common Features
  • Usage Example