Decision Tree-based Machine Learning Algorithms to Classify Rice Plant Diseases: A Recent Study

Abstract

Rice is one of the most important foods for humans on earth. India and China are two of the world's most rice-dependent countries. The yield of this crop is determined by a number of factors, including soil, water supply, pesticides used, time period, and disease infection. Rice Plant Disease (RPD) is one of the most significant factors affecting rice quantity and quality.  Farmers face a constant challenge in determining the form of rice plant disease and taking timely corrective action against it. Although the rice plant is susceptible to a variety of diseases, the most common are Bacterial Leaf Blight (BLB), Brown Spot (BS), and Leaf Smut (LS).Since the infected leaf must be processed by the human eye, identifying this disease is extremely difficult. To define and classify the RPD, we used machine learning techniques in this chapter. We used the UCI Machine Learning repository to gather data on contaminated rice plants. The data collection contains 120 images of contaminated rice plants, with 40 BLB images, 40 BS images, and 40 LS images. RandomForest, REPTree, and J48 are decision tree-based machine learning algorithms used in the experiments.  We used ColorLayoutFilter, which is provided by WEKA, to extract numerical features from the infected images. The experimental analysis makes use of 65% of the data for training and 35% of the data for testing. The Random Forest algorithm performs exceptionally well in predicting RPD, according to the experiments.