Perbandingan Neural Network MLP, KNN, dan Decision Tree untuk Klasifikasi Penyakit Diabetes

Abstract

Diabetes is one of the diseases that has received global attention due to its extensive impact on public health. Most people with diabetes are unaware that they are suffering from this condition, this situation emphasizes the need for improved understanding and more effective treatment of this disease. In an effort to address these challenges, this study compares three machine learning algorithms for diabetes classification, the three algorithms are: Multi-Layer Perceptron (MLP), K-Nearest Neighbor (KNN), and Decision Tree. Data from the Diabetes Dataset used to train and test these models will go through preprocessing first starting from data cleaning, encoding because there is string data, data distribution analysis where in this study using under sampling to equalize data and normalization using min-max normalization, Evaluation results using Confusion Matrix and Classification Report which contains precision, recall, and f1-score the results of this evaluation show that the Neural Network MLP model achieves the highest accuracy of 90.48%, followed by KNN with 88.15% accuracy, and Decision Tree with 87.24% accuracy. These findings provide important insights in selecting the optimal model for diabetes prediction applications.

Keywords: Diabetes, Machine Learning, Neural Network MLP, KNN, Decision Tree