<![CDATA[The author's previous research compared XGBoost and Linear Regression for estimating building heating and cooling loads, from the results of the study obtained the smallest RMSE using XGBoost of 1 and RMSE (Root Mean Square Error) is one of the benchmarks for the accuracy of a study with a regression dataset, so in this study the author carried out optimization in estimating building heating and cooling loads using hyperparameters with XGBoost, as a continuation of previous research to obtain a better RMSE value, using the same dataset energy efficiency from the UCI website. The dataset division in the research that the author is currently conducting is 80% training data and 20% for testing data, the author also changed the random state to 42 as well as the choice of value combinations in learning_rate, n_estimator, max_depth, subsample, colsample_bytree and min_child_weight. The author uses the python programming language to get the best combination and in this study, both heating and cooling loads, theauthor got the best combination at a learning rate of 0.05, n_estimator 1000, max depth 10, subsample 0.8, colsample bytree 0.5 and min child weight 3. For the heating load, the RMSE value was 0.3575 and the R² score was 0.9988, while for the cooling load, the RMSE value was 0.7978 with an R² score of 0.9931, so it can be concluded that the RMSE value is much better than previous research.]]>
Copyrights © 2026
