IMPLEMENTASI ALGORITMA RANDOM FOREST PADA KLASIFIKASI DATA PEMANFAATAN TANAH PROGRAM IP4T

The IP4T (Inventory of Land Tenure, Ownership, Use, and Utilization) program is a national strategic initiative aimed at managing and organizing land data systematically. However, the complexity and large volume of land utilization data pose challenges for manual Classification. This study aims to implement the Random forest Classifier algorithm to classify land utilization data in support of effective decision-making within the 2024 IP4T program. The research data were obtained from the National Land Agency (BPN) of Sukabumi Regency, consisting of attributes such as land tenure, land ownership, land utilization, land use, land area, and TOL (Land Object for Land Reform) potential. The methods applied include data preprocessing, transformation using One-Hot Encoding, data splitting into training and testing sets, handling class imbalance using SMOTE (Synthetic Minority Over-sampling Technique), and training and evaluating the Random forest model. The evaluation results showed a model Accuracy of 94.8%, with high Precision, Recall, and F1-score values for each target class: Access Reform: Precision 0.95, Recall 0.94, F1-score 0.94 TORA Potential: Precision 0.94, Recall 0.95, F1-score 0.94 Dispute: Precision 0.94, Recall 0.96, F1-score
0.95 Asset Legalization: Precision 0.94, Recall 0.94, F1-score 0.94. The model was then integrated into a web interface using the Streamlit framework, which features three main components: Home, Classification Data Information, and TOL Potential Prediction. The results indicate that the Random forest algorithm is effective in classifying land utilization data, and the developed system provides visual inSIGhts to support agrarian reform through the IP4T program.
Keywords: Random forest, IP4T, Land Utilization Data Classification.