Optimalisasi Regresi Logistik Menggunakan Algoritma Genetika Pada Data Klasifikasi

Abdurrahman Salim (Politeknik Negeri Jember)
Muhammad Rijal Alfian (Universitas Teknologi Mataram, Nusa Tenggara Barat, Indonesia)

Jurnal Teknologi Informasi dan Terapan (J-TIT)
Vol 6 No 2 (2019)

Abstract— Classification on large of data, and with a variety of features or attributes often makes the law accuracy. It required a method that has immunity in such diverse data types. One of method is Logistic Regression method. Logistic Regression is one of classification method, if response variable has binary characteristic and there are many predictor variable such as combination of category and continue.Methd of Logistic Regression requires a stage selection independent variable in improving the model accuration. So it takes a good method in fixing the deficiency is Genetic Algorithm (GA). This method is an iterative method to get global optimum. The results of the classification accuracy of Logistic Regression in the case of septictank data in East Surabaya with 11 independent variables and binary dependent variable is Logistic Regression accuracy of 54.55%. However when selected with GA, the classification accuracy of Binary Logistic Regression is 90.91%.