A boosting ensemble learning based hybrid light gradient boosting machine and extreme gradient boosting model for predicting house prices

The implementation of tree-ensemble models has become increasingly essential in solving classification and prediction problems. Boosting ensemble techniques have been widely used as individual machine learning algorithms predicting house prices. One the is LGBM algorithm that employs leaf wise growth strategy, reduces loss improves accuracy during training which results overfitting. However, XGBoost uses level strategy takes time to compute resulting higher computation time. Nevertheless, a regularization parameter, implements column sampling weight reduction on new trees combats This study focuses developing hybrid model order prevent overfitting through minimizing variance whilst improving accuracy. Bayesian hyperparameter optimization technique implemented base learners find best combination hyperparameters. resulted reduced (overfitting) since parameter values were optimized. compared LGBM, XGBoost, Adaboost GBM evaluate its performance giving accurate price predictions using MSE, MAE MAPE evaluation metrics. outperformed other with 0.193, 0.285, 0.156 respectively.