Perbandingan Metode Fuzzy Dengan Regresi Linear Berganda Dalam Peramalan Jumlah Produksi (Studi Kasus: Produksi Kelapa Sawit di PT. Perkebunan Nusantara III (PERSERO) Medan Tahun 2011- 2012)

Abstract

This study is shown how to compare the result of prediction by using Fuzzy Sets and multiple linear regression. In this study, production of palm oil is used as output or dependent variable (Y), and the manuring, worker, and avarage of rainfall are used as output or independent variable X1, X2, and X3. The manuring variable (X1) consist of three fuzzy sets : minimum, standard, maximum. For worker variable (X2) consist of three fuzzy sets : minimum, normal, maximum. For avarage of rainfall (X3) consist of three fuzzy sets : low, standard, high. Meanwhile, production of palm oil consist of three fuzzy sets : decrease, permanent, increase. In this study, fuzzy use 16 fuzzy rules. The solution of fuzzy logic use fuzzy-Mamdani Method. Multiple regression linear analysis use least squares method as the solution. By showing the avarage of error relative from both of methods, which for fuzzy set is 0,20748 atau 20,748% and for linear regression is 0,09383 atau 9,383%. It’s value shows that the avarage of error relative from linear regression is smaller than fuzzy set. So for the case where input and output in this study, found the conclusion that prediction with multiple regression linear analysis is better than using fuzzy logic.