An Ensemble Based Clustering and Classification Framework for Prediction of Agricultural Crop Yield

Abstract

Agriculture, a critical sector in India's economy, is deeply influenced by various factors such as climate, soil quality, irrigation, and economic conditions. Accurate prediction of crop yield is essential for both farmers and businesses operating in the agricultural supply chain. Historical data on crop yields can inform decision-making, reduce risks, and enhance the efficiency of supply chain operations, including the scheduling of production and the marketing of agricultural products. However, existing methods for predicting crop yield often face challenges in accuracy and generalizability.This study proposes an ensemble-based clustering and classification framework to improve the prediction of agricultural crop yield. The framework includes three main components: (i) pre-processing of agricultural datasets, (ii) ensemble clustering using two optimization algorithms—Enhanced Artificial Bee Colony Optimization (EABCO) and Shuffled Frog Leaping Algorithm (SFLA), and (i) classification of clustered datasets using Enhanced K-Nearest Neighbor (KNN) classification. The proposed method aims to reduce the limitations farmers face in selecting suitable crops for specific regions and to enhance the accuracy of crop yield predictions by identifying key indicators of agricultural field heterogeneity.The framework was implemented in a Java environment and tested on datasets related to paddy crop yield. The performance of the proposed method was evaluated using accuracy, precision, recall, F-measure, and percentage error. The results demonstrate that the proposed EABCO-SFLA clustering ensemble, combined with Enhanced KNN classification, achieves higher accuracy (92.7%), improved precision (91.2%), better recall (90.6%), and a reduced percentage error (5.3%) compared to existing classifiers. This approach provides a more reliable prediction of crop yield, enabling better decision-making in agriculture and supply chain management.