<![CDATA[This study aims to implement the Random Forest Regression algorithm to predict maize yields in North Luwu Regency, South Sulawesi. Historical data from 2020–2024 were used, with variables including planting area, harvested area, productivity, and production. The study follows the CRISP-DM methodology, which includes the stages of business understanding, data understanding, data preparation, modeling, evaluation, and deployment. Modeling was carried out using Google Colab with GPU/TPU support and Google Drive integration. The dataset was divided into 80% training data and 20% testing data. The model was developed with parameters such as max_depth = 10 and n_estimators = 200. Evaluation results indicated excellent performance, with an R² value of 0.9974, RMSE of 42.98 (4.02%), MAE of 26.188 (2.45%), and MAPE of 7.12%, all of which fall under the “excellent” category. The trained model was then integrated into a website to facilitate users in predicting maize production based on input variables. Questionnaire results from 10 respondents showed a very high satisfaction level, with an average score of 96.6%, classified as excellent.]]>
Copyrights © 2025
