<![CDATA[This study aims to develop a forecasting model for the number of tourists at five tourist destinations in Batang Regency using historical tourist visit data and external factors, as well as to implement the best model into an application. The methods used are Random Forest Regressor, XGBoost Regressor, GRU, and LSTM, with stages based on Knowledge Discovery in Databases (KDD), feature selection, hyperparameter tuning, and evaluation using MAPE, RMSE, MAE, and R2. The results show that the best model was obtained from XGB-4, with an average MAPE of 4.96%, RMSE of 526.93, MAE of 428.07, and R2 of 0.972. The features that significantly affect the model include Exponantial Moving Average (EMA), trend, lagging, volatility, temperature max, and rain. The best model was implemented in an application called LARAS BATANG, which stands for Layanan Ramalan Wisata Batang. The application displays tourism forecasting results with several destination options and forecasting periods. In addition, the application can show a forecasting prototype that includes a summary, a time series line chart, a donut chart of visitor data proportions, and a feature importance bar chart.]]>
Copyrights © 2026
