A Framework System for Plant Leaf Disease Detection using K-Nearest Neighbours and comparison of its features with Naive Bayes Classification

Abstract

Aim: To perform leaf disease detection using K-nearest neighbour (KNN) algorithm and comparing its accuracy with Naive Bayes(NB) algorithm. Methods: In this proposed work, the plant leaf disease detection has been carried out using machine learning algorithms such as KNN (N=10) and NB (N=10) and the accuracy was determined for the same. Results: From the implemented experiment, the NB algorithm’s leaf disease accuracy is significantly (0.604) appeared to be better than the KNN algorithm. The accuracy of leaf disease was compared and the NB algorithm’s accuracy appears to be higher 91% than KNN algorithm accuracy 83%. Conclusion: The result shows that NB algorithm’s accuracy was better than KNN algorithm accuracy for leaf disease detection.