Comparison of Fuzzy Logic and Multiple Linear Regression in Forecasting Rice Production in Toba District

Abstract

The purpose of this study is to forecast rice production in Toba Regency in the future. Factors affecting rice production studied are land area, subsidized fertilizers, rice pest populations, average rainfall on rainy days. The methods used in this study are fuzzy logic methods and linear regression. This study compared more accurate forecasting results based on errors from both methods. The data used is secondary data from the Toba Regency Agriculture Office. The data studied were in the form of data on rice production, land area, subsidized fertilizers, pest populations, average rainfall and rainy days in 2016-2021. This research was conducted with the help of matlab toolbox fuzzy and SPSS software. From the forecasting results obtained, the multiple linear regression method is closer to the actual result of 878,428.07 Tons compared to using fuzzy logic of 1,032,300 Tons. Based on the error standards of the two methods, it can be concluded that the multiple linear regression method has a more accurate result with an error of 3.11% compared to the fuzzy logic method with an error of 22.05%.
 Keywords : Rice, Production, Fuzzy Logic, Linear Regression, Standard Error.