Performance Evaluation of Machine Learning Algorithms for Fruit Disease Identification

Main Article Content

Dr. Santosh Jagtap

Abstract

Fruit diseases have long been recognized as one of the major causes of reduced agricultural productivity and post-harvest losses (Patil et al., 2019; Zhang et al., 2020). Early and precise detection of such diseases plays a crucial role in maintaining fruit quality and ensuring sustainable farming practices (Singh et al., 2018). This study presents a machine learning-based framework for the identification of fruit diseases using image datasets. Four algorithmsConvolutional Neural Network (CNN), K-Nearest Neighbor (KNN), Decision Tree (DT), and Support Vector Machine (SVM) and the Hybrid-CNN-SVM model were applied and their performance compared using standard metrics such as accuracy, precision, recall, and F1-score (Kaur and Sood, 2019; Sharma et al., 2020). To improve detection performance, a Hybrid-CNN-SVM model was proposed, integrating CNN’s feature extraction capabilities with SVM’s strong classification performance (Jadhav and Patil, 2020). The hybrid model achieved an overall accuracy of 98.3%, outperforming the individual algorithms. The results confirm that the combination of deep feature extraction and machine learning classification offers higher accuracy and reliability for fruit disease recognition. This approach can support early disease detection, reduce yield losses, and contribute to smarter agricultural management systems (Li et al., 2019; Meena et al., 2020).

Article Details

Section
Articles