Perbandingan Algoritma Naive Bayes Berbasis Feature Selection Gain Ratio dengan Naive Bayes Kovensional dalam Prediksi Komplikasi Hipertensi
Abstract
High blood pressure is a significant public health problem with a high prevalence in the Indo-nesian population. 2018 Riset Kesehatan Dasar (Riskesdas) data shows a prevalence of hyperten-sion of 34.1% in individuals aged 18 years and over, with the highest figure in South Kalimantan and the lowest in Papua. Complications arising from hypertension can have serious impacts on organs such as the brain, eyes, heart and kidneys. The Naive Bayes algorithm is generally used in disease prediction, but the Naive Bayes algorithm has problems when selecting attributes, because Naive Bayes itself is a statistical classification method that is only based on Bayes' Theorem so it can only be used with the aim of predicting the probability of membership in a group or class. So attribute weighting is needed to increase accuracy more effectively. This research introduces Gain Ratio as an attribute weighting method to increase the accuracy of Naive Bayes. The aim of this study was to compare conventional Naive Bayes with Naive Bayes Gain Ratio in predicting complications of high blood pressure. The research results show that feature selection with gain ratio weighting can increase the accuracy of naive Bayes classification, with an average increase in accuracy of 20% compared to naive Bayes without feature selection. The precision value increased by 21% in the naive Bayes gain ratio algorithm for the kidney failure class, an increase of 3% in the heart class, and an increase of 31% in the stroke class, for the recall value the naive Bayes gain ratio increased by 35% in the heart class while in the kidney failure and stroke classes did not increase the recall value.
Downloads
References
[2] Kemenkes, “Hipertensi Penyakit Paling Banyak Diidap Masyarakat,” 17 Mei 2019. [Online]. Available: https://www.kemkes.go.id/article/view/19051700002/hipertensi-penyakit-paling-banyak-diidap-masyarakat.html.
[3] A. Adinata, “Pengembangan Tensimeter Digital Bayesian Berbasis Arduino Untuk Prediksi Komplikasi,” Universitas Pendidikan Ganesha, Singaraja, 2019.
[4] R. Rizky, S. Z. Hakim dan S. Lili, “Sistem Pakar Deteksi Penyakit Hipertensi Dan Upaya Pencegahannya Menggunakan Metode Naive Bayes Pada RSUD Pandeglang Banten,” Jutis (Jurnal Teknik Informatika), 2019.
[5] M. M. Santoni, N. Chamidah dan N. Matondang, “Prediksi Hipertensi Menggunakan Decision Tree, Naïve,” Techno.COM, 2020.
[6] N. A. Siagian, “Pembobotan Fitur Dataset Menggunakan Gain Ratio Guna Meningkatkan Akurasi Metode Naïve Bayesian Classifier,” Universitas Sumatera Utara, Medan, 2020.
[7] F. Ghasemi, B. S. Neysiani dan N. Nematbakhsh, “Feature Selection in Pre-Diagnosis Heart Coronary Artery Disease Detection: A heuristic approach for feature selection based on Information Gain Ratio and Gini Index,” dalam International Conference on Web Research (ICWR), 2020.
[8] I. Santiko dan I. Honggo, “Naive Bayes Algorithm Using Selection of Correlation Based Featured Selections Features for Chronic Diagnosis Disease,” International Journal of Informatics and Information Systems, 2019.
[9] A. N. Desai, “High Blood Pressure,” JAMA, 2020.
[10] E. Prasetyo, Data Mining Pengelompokkan Data untuk Informasi dan Evaluasi, Surabaya: Duta Media, 2017.
[11] D. Jollyta, Konsep Data Mining dan Penerapan, Deepublish, 2020.
[12] M. I. Prasetiyowati, N. U. Maulidevi dan K. Surendro, “Determining Threshold Value on Information Gain Feature Selection to Increase Speed and Prediction Accuracy of Random Forest,” Research Square, 2020.
Copyright (c) 2024 I Made Arya Adinata Dwija Putra, I Made Gede Sunarya, I Gede Aris Gunadi
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
