Prediction of Potato Late Blight Disease Based Upon Weather Parameters Using Artificial Neural Network Approach

Late Blight in potato is caused by the oomycete pathogen Phytophthora infestans. It is known as community disease due to its ability to spread rapidly. It has caused huge economic losses. International Potato Center(CIP) has made a global estimate of Late Blight damage in developing countries based on an average production loss of 15%. The proposed paper tends to find a relationship between weather parameters and manifestation of Late Blight using machine learning approach. The disease manifestation is predicted using artificial neural network and classified on the basis of disease severity. The architecture of network is of feedforward type and standard backpropogation algorithm is used for learning. The weather parameters considered in the study are-maximum temperature, minimum temperature, maximum humidity, minimum humidity, and rainfall. The database used in the thesis is compiled from the data used for AICRP on Potato. Being the maiden venture in the domain of crop disease prediction using ANN, the basic spade work of the data organization and representation was performed. Later on a suitable architecture and activation function were identified. The prediction accuracy of 90. 9% was achieved.