INTEGRATING STACKING ENSEMBLE LEARNING FOR SOIL ANALYSIS AND CROP PREDICTION USING OPTIMIZED FEATURE SELECTION APPROACH

Abstract

Accurate and efficient crop prediction is crucial for optimizing agricultural productivity and supporting sustainable farming practices. In precision agriculture, accurate crop recommendations are vital for improving agricultural productivity and sustainability. This research proposes an advanced crop recommendation system that integrates Stacking Ensemble Learning with an Optimized Feature Selection approach for enhanced soil analysis and crop prediction. Additionally, an optimized feature selection technique is applied to identify the most relevant features from large agricultural datasets, enhancing model performance by eliminating irrelevant or redundant data. This work utilizes soil properties, climate conditions, and historical crop data to recommend the most suitable crops for specific environments. In this paper, the feature selection techniques namely Recursive Feature Elimination (RFE) and Boruta are used to select the features. Integrating Stacking Ensemble Learning (ISEL) technique consists of k-Nearest Neighbor (KNN), Support Vector Machine (SVM), Naïve Bayes (NB), Random Forest (RF) and Decision Tree (DT) algorithms to classify the classes. The Meta classifier SVM is used to predict and recommend the suitable crops to the farmers.