Crop Prediction Based On Characteristics of the Agricultural Environment Using Various Feature Selection Techniques and Classifiers

Abstract

Crop prediction is a pivotal aspect of modern agriculture, enabling farmers and stakeholders to make informed decisions regarding crop selection and resource allocation. The integration of machine learning techniques has revolutionized this domain by facilitating the analysis of complex agricultural datasets. This study explores the efficacy of various feature selection methods and classifiers in predicting suitable crops based on environmental characteristics. By employing techniques such as Recursive Feature Elimination (RFE), Principal Component Analysis (PCA), and Information Gain for feature selection, and classifiers including Random Forest, Support Vector Machine (SVM), and Naïve Bayes, we aim to enhance prediction accuracy. The dataset comprises diverse environmental parameters like soil type, pH, temperature, and rainfall. Experimental results indicate that the combination of RFE and Random Forest yields the highest accuracy, underscoring the significance of optimal feature selection in crop prediction models. This research contributes to the development of intelligent agricultural systems that support sustainable farming practices and efficient resource utilization.