Klasifikasi Penyakit Jantung Dengan Neural Network dan Seleksi Fitur Chi-Square

Abstract

This study aims to develop a heart disease classification model using Neural Networks (NN) combined with the Chi-Squared feature selection method. Heart disease is one of the leading causes of death worldwide, making early detection crucial. The dataset used in this research was obtained from the Kaggle platform and underwent preprocessing steps, including outlier handling and data normalization. Feature selection was performed using the Chi-Squared method to identify the most relevant features. Several feature selection scenarios and NN configurations were tested. The results indicate that using NN with Chi-Squared feature selection can improve prediction accuracy up to 86%, depending on the number of features and hyperparameter configurations used. In conclusion, this method combination effectively enhances heart disease detection accuracy.