Seleksi Fitur dengan Menggunakan Metode Information Gain pada Algoritma Logistic Regression pada Penyakit Diabetes

Abstract

Diabetes is one of the significant global health problems with widespread impacts on individual quality of life and a significant economic burden on healthcare systems. In an effort to improve early diagnosis and understanding of factors influencing this disease, the use of data analysis techniques has become increasingly important. One approach used is the application of logistic regression algorithms, which provide information on the probability of diabetes occurrence based on independent variables. In this study, the use of Information Gain-based feature selection methods is explored to enhance the performance of logistic regression algorithms in identifying risk factors for diabetes. Information Gain method is employed to evaluate the relevance of variables to the target class, i.e., the presence or absence of diabetes. In the experimental process, a dataset consisting of clinical attributes such as age, body mass index (BMI), blood pressure, and several other biochemical parameters is used. The experimental results indicate that the use of Information Gain method for feature selection can improve the performance of logistic regression models in predicting the presence of diabetes. By reducing the dimensionality of irrelevant attributes, the resulting model tends to have higher accuracy and can identify more significant risk factors. This highlights the potential of Information Gain-based feature selection methods in enhancing the efficiency and effectiveness of predictive analysis in diabetes.