Machine Learning Techniques for Disease Prediction in Rice Plant

It is essential to identify diseases in plants to increase crop production and yield. Therefore, prediction of these diseases at an early stage is important to avoid or minimize crop losses. In this paper we have taken the use case of rice crop and predicting diseases namely leaf smut, brown spots and bacterial leaf blight using machine learning techniques. Labeled image dataset of these three diseases in rice plants is used and several machine learning algorithms (DT, RF, KNN, GBM, LR, NB, and MobileNet) are applied on it. To compare performance of different models, calculations and analysis of confusion matrix, recall, precision and f1-score have been used.