Prediction of Wheat Fusarium Head Blight Severity by Using Random Forest

Rapid increase of the world population growth and the demand for food security makes increasing yield as an essential strategy for solving the food supply problem. What is more, because of the restrictions in increasing crop cultivation areas and the decrease in some crops such as wheat in Iran, increasing the yield potential can be an effective way to respond to this requirement. Fusarium Head Blight (FHB) is one of the most important wheat diseases and for prediction FHB some methods have already been developed in the USA, Canada, Argentina and Brazil. As there is no model for predicting FHB in Iran, in this study, a method for predicting severity of FHB based on spatial analysis and using environmental parameters and meteorological data was developed for the Moghan, which is in the northwest of Iran. An Internet of Things (IoT) network was established in the study area for measurement of environmental data, including relative humidity, rainfall and air temperature for evaluating the developed model. Random Forests (RF) and extracted indices were used for predicting FHB severity and calculating the relative importance of the indices. We evaluated FHB for the period of 1389 to 1396 and the results show the effectiveness of the developed model and the capability of IoT and spatial analysis for predicting FHB.