A Hybrid Approach for Prediction and Classification in Rice Plant Disease Using Machine Learning Algorithm

Abstract

Agriculture has a crucial part in the entire life of a given economy and also provides employment opportunities in the world. In agriculture, Rice is a staple food for much more than half of the total population and one of the most significant food crops in the world. But the quality and quantity of rice cultivation decreases due to plant diseases. Organisms like bacteria, virus, and fungi are the major causes of plant diseases.  Manually observing the field on daily basis is not likely at all the times by the farmers to protect them from infection as well as for proper irrigation. In order to decrease the damage to crops because of diseases and escalate productivity Machine Learning (ML) techniques have been industrialized, which automates the recognition of crop disease. The aim of this paper is to review various ML algorithms such as Decision Tree, KNN, Random forest, ANN, Support vector machine which focuses on the prediction of plant diseases, to maximize crop yield. After comparing various techniques it is proposed that the combination of Decision tree and Convolutional Neural Network (CNN) algorithm improves the accuracy and prediction of the disease at an early stage.