Implement Gradient Boosting Regressor with Decision Trees in R

[iampratik13]

This PR introduces a fully functional Gradient Boosting Regressor implementation in R, designed for educational and practical purposes. Gradient Boosting is a sequential ensemble learning method where each model iteratively corrects the errors of previous models, making it a powerful technique for regression problems.

Algorithm Complexity:
• Time complexity: O(n_trees × n_samples × log(n_samples))
• Space complexity: O(n_trees × tree_size)

[Copilot]

Pull Request Overview

This PR implements a complete Gradient Boosting Regressor algorithm in R using object-oriented programming with R6 classes. The implementation includes decision trees as weak learners and demonstrates the sequential ensemble learning approach where each model corrects errors from previous models.

• Implements three R6 classes: DecisionTreeNode, RegressionTree, and GradientBoostingRegressor
• Provides comprehensive demonstration with synthetic datasets and hyperparameter comparison
• Includes validation, prediction methods, and feature importance calculation