Analysis of the Stacking Ensemble Learning Model of Categorical Boosting and Naïve Bayes Algorithms for Crop Selection Based on Soil Characteristics

This study aims to develop a machine learning model for selecting crop types based on soil characteristics, using the Categorical Boosting and Naïve Bayes algorithms as base learners. Next, an ensemble learning technique using a stacking approach was applied to improve the performance of the base model that was built. This was done to analyze and compare the performance results of each ensemble model that was carried out. Model performance was evaluated using evaluation metrics including precision, recall, f1-score, and accuracy. The results of this study indicate that the stacking ensemble model with Random Forest as the meta learner can provide better performance compared to other ensemble models. This model achieved a precision of 98.85337%, a recall of 99.84848%, an F1-score of 99.84844%, an accuracy of 99.84848%, and a model training time of 78.61110 seconds. Based on these results, this study is expected to provide tangible contributions and new knowledge in plant selection classification based on soil characteristics, thereby aiding in the precise and efficient determination of suitable plant types.