IMPLEMENTASI ALGORITMA REGRESI LOGISTIK MENGGUNAKAN MACHINE LEARNING UNTUK MEMPREDIKSI DIABETES

Diabetes needs to be predicted accurately because it is a serious social disease that can affect a large number of people, cause complications, involve high costs, and improve the condition of diabetes, especially in children and young people. Therefore, to reduce the harmful effects of diabetes on people and society as a whole, a better understanding of risk factors, prevention, and treatment of diabetes is essential. The aim of this study is to integrate logistical regression algorithms to predict diabetes or not in the data obtained from Kaggle.com, which has a data set of 768. Based on previous research, the authors use logistics regression algorithms to create intelligent systems that can predict or not diabetes. The data itself consists of 768 records with several medical predictor variables or attributes (pregnancies, glucose, blood pressure, skin thickness, insulin, BMI/body time index, diabetes pedigree function, age, and outcome). The results of this study show that using evaluation tools such as cross-validation and a confusion matrix can make the results fairly good and accurate. For training data, the results were 77.1%, and the results from testing data produced 74%. While accuracy results using Python obtained 77.54% for the results of training data, the results on testing data were 78.57%. The ultimate goal of this research is to develop a logistic regression model to predict diabetes data that can contribute and be a recommendation for helping the public, even medical personnel, identify the factors that cause diabetes so that the public can prevent the occurrence of diabetes.

Keywords: predictions, diabetes, logistical regression algorithms, cross-validation, confusion matrix, Python.