Crop Disease Prediction Using Multiple Linear Regression Modelling

Agriculture is a key player in the economic growth and sustainability of Small Island Developing States like Mauritius. However, during the past decade, climatic variations in Mauritius have caused economically important crops such as onion, potato, and tomato, to become more vulnerable to diseases, representing a severe threat to food security. Diseases caused by fungal microorganisms are expected to increase since climatic factors such as temperature, humidity, rainfall, and radiation directly affect fungal infection, growth and spread. Thus, mechanisms for efficient crop disease prediction based on climatic data are more than ever needed. Several applications exist for climatic data prediction but none of them have been adapted for Mauritius. In this paper, we have recorded climatic variables and collected crop disease incidence data from regions in different agro-climatic zones of Mauritius over a period of two years. We have subsequently developed an android mobile application implemented in Mauritian Kreol which applies Multiple Linear Regression modelling to the collected data for crop disease prediction. The application also acts as an “information window” which provides planters with a catalogue of diseases to aid diagnosis, cultivation tips, and appropriate treatments. Our findings indicate that the models developed were able to fit the collected data with fair R-squared values ranging from 0.14–0.28. We anticipate that our application can help farmers in Mauritius and in other similar countries to better forecast incidents of crop diseases and take remedial actions accordingly.