METODE CLASSIFICATION-DECISSION TREE DAN ALGORITMA SELF ORGANIZING MAP UNTUK MEMPREDIKSI PENYAKIT DIABETES DENGAN MENGGUNAKAN APLIKASI RAPID MINER

Abstract

Diabetes mellitus is a chronic disease that requires early detection to prevent serious complications. This study used Decision Tree and Self-Organizing Map (SOM) algorithms to predict diabetes risk based on a dataset of patients ' clinical symptoms. The Dataset consisted of 520 data with 16 attributes, including age, sex, and symptoms such as polyuria, polydipsia, and obesity. The Decision Tree method is applied to build a classification model that is able to identify the main risk factors, while SOM is used to Group data based on similarity patterns and provide informative visualization. The analysis was carried out using the RapidMiner application to simplify the process of data preprocessing, modeling, and evaluation of results. The results showed that the Decision Tree provides high accuracy with a clear interpretation of the results, while SOM produces clusters that represent the patient's symptom patterns. The combination of these two methods is effective in supporting early detection of diabetes and aiding in better medical decision making