Optimizing Crop Recommendation Systems Using Machine Learning Algorithms

Abstract

Agriculture remains a vital sector, providing essential sustenance and livelihoods worldwide. However, selecting the optimal crops for specific fields remains a significant challenge, impacting both yields and income. The decline in land fertility exacerbates this issue, necessitating innovative solutions for crop recommendation. This study explores the integration of advanced technologies, including artificial intelligence (AI) and machine learning (ML), to enhance crop recommendation systems. Specifically, it focuses on developing a smart crop recommendation model using optimization algorithms such as Particle Swarm Optimization (PSO) and Shuffled Frog-Leaping Algorithm (SFLA) to optimize Support Vector Machines (SVM). The model aims to predict suitable crops based on soil properties like pH, humidity, temperature, and nutrient levels (N, P, K). Extensive literature review highlights various ML approaches previously employed, underscoring the necessity for more comprehensive and accurate systems. Data for this study was sourced from the Crop Recommendation Dataset on Kaggle. The dataset underwent preprocessing to enhance its quality and facilitate effective model training. The study employed SVMs for initial model training, followed by optimization using PSO and SFLA. Performance metrics including accuracy, precision, recall, specificity, and F1 score were utilized to evaluate the models. Post optimization results demonstrated significant improvements, with PSO achieving an accuracy of 98.64% and SFLA 97.27%, indicating the potential of these approaches in revolutionizing crop recommendation systems. Future recommendations include leveraging advanced ML techniques like deep learning and reinforcement learning to enhance crop recommendation further