APPLYING TREE BASED MODEL FOR CROP RECOMMENDATION SYSTEM BASED ON SOIL PARAMETERS AND WEATHER CONDITIONS

The massive population in Indonesia needs to be supported by various sectors so that the population's needs are met. One of these sectors is agriculture. The problems are unpredictable climate change and weather and changes in land use from previously agricultural land to housing. In addition, plant quality is also influenced by soil quality and other abiotic factors, comprising rainfall, temperature, and humidity. Plant quality affects the increase in crop yields. A plant recommendation system based on plant parameters must help farmers determine the best plants according to agricultural land conditions. The recommended plants to be used include mango, cotton, rice, mungbeans, and apple. This work aims to create a plant recommendation system utilizing criteria related to plant requirements through a machine learning methodology. The stages in this study start with data collection, preprocessing, partitioning, modelling, performance evaluation, and a recommender system. This study’s results indicate that the Random Forest method achieved the best accuracy at 0.9981, followed by XGBoost at 0.9909 and Decision Tree at 0.9873. The system provided recommendations for plant types based on user input