Kajian Analisis Regresi Linier Bergerombol dengan Pendekatan Algoritma Genetika

Abstract

Clustered Linear Regression (CLR) analysis with Genetic Algorithms (GA) approach is able to find solutions to the data clustering based on the adjusted Rsquare (R2 adj) value, but have not been able to determine the optimum clusters because of the larger number of clusters the R2 adj value also higher. This analysis is able to determine cluster for an observation so that it can improve the ability of linear regression analysis. If compared to classical linear regression analysis, this analysis is able to produce better estimation of parameters and R2 adj value. CLR analysis with GA approach program can be accessed at http://bit.ly/CLRwGA