Machine Learning Model of Plant Disease Prediction: Cotton Leaf Curl Virus (CLCV) on Cotton Crop

Abstract

The attack of various diseases is one of the major issues in agriculture, and it has a significant impact on the productivity of agriculture. The correlations between the plant disease life cycle and environmental conditions can be successfully used for the prediction and early warnings of the occurrence of diseases. This study proposes a machine-learning approach for the prediction of disease attacks based on environmental conditions. The proposed model is based on the crop field Internet of Things (IoT) based directly sensed temperature, humidity, and rainfall to predict the monthly occurrence of the Cotton Leaf Curl Virus (CLCV) disease on cotton crops. Moreover, the proposed regression model enables the prediction of the CLCV disease attack with a multiple co-efficient of co-relation value of 0.88 and a coefficient of determination (R2) of 0.82 from the prevailing temperature, humidity, and rainfall environmental conditions. The field evaluation reveals that the predicted accuracy of the proposed solution is 85%.