Ensemble-Based Predictive Model for Crop Recommendation

ABSTRACT:

Farmers frequently face the challenge of choosing the best crops for their specific locations to improve production and profitability. This paper developed a prediction model that uses stacking ensemble approaches to address this issue. The study involves the use of a correlation matrix to analyze secondary data obtained from Kaggle. The data was utilized to train the developed model incorporating machine learning algorithms, such as Decision Tree, Support Vector Machine (SVM), Random Forest, Naive Bayes model, XGBoost, Logistic Regression (LR), and K-Nearest Neighbors (KNN). The Random Forest was used as a meta-model in this study. The proposed approach demonstrates a very high level of accuracy, achieving 99.8%, surpassing the performance of previous investigations. This study demonstrates the potential of predictive modeling in providing farmers with cutting-edge tools to make smarter decisions, leading to significant improvements in productivity and profitability.

Keywords: Ensemble Learning, Predictive Model, Crop Recommendation, Stacking, Machine Learning.