multi factor linear regression

Intelligent systems have received considerable attention as a modern modeling methods in recent years. These models is used for prediction and classification in situations where the classic statistical models are not able due to their constraints. This study is aimed to compare the ability of ANFIS and multi factor linear regression models for predicting density of all growing stages of Sunn pest. The data population fluctuation of Sunn pest in the years 2015 and 2016 on a farm with an area of one hectar in chadegan city was obtained. Predictor variables including variables sampling date, average temperature, average relative humidity, wind speed, wind direction, rainfall, height from sea level and degree- day were processed as input data to achive an output of number of developmental stages as response variable. In the ANFIS model, 70% of the data was assigned to training and 30% for validation. After network training and assessment of the best structure according to type, number of membership function and related rules with the use of MATLAB software, the appropriate model was selected based on statistical indices of, root mean square error (RMSE) and coefficient of determination (R2). After sensivity analysis the results showed that ANFIS method (RMSE= 0.051, R2= 0.97) had higher accuracy than multi linear regression (RMSE= 0.26, R2= 0.47) and better predicts the population fluctuation of Sunn pest Eurygaster integriceps.