Indonesian Journal of Applied Informatics
Vol 10, No 1 (2025)

Perbandingan Algoritma Machine Learning untuk Klasifikasi Risiko Penyakit Paru Berdasarkan Data Diagnostik Pasien

Alwan Kapi Muntaha (Universitas Bina Sarana Informatika)
Kevin Dwi Satria (Universitas Bina Sarana Informatika)
Desiana Nuranudin Putri (Universitas Bina Sarana Informatika)
Sumanto Sumanto (Universitas Bina Sarana Informatika)
Roida Pakpahan (Universitas Bina Sarana Informatika)
Imam Budiawan (Universitas Bina Sarana Informatika)



Article Info

Publish Date
19 Jan 2026

Abstract

Abstrak : Penyakit paru-paru termasuk salah satu faktor utama penyebab tingginya angka kematian di seluruh dunia. Kondisi ini terjadi karena penyakit paru-paru sering kali sulit terdeteksi pada tahap awal akibat gejalanya yang tidak spesifik. Perkembangan teknologi machine learning memberikan peluang untuk membantu proses diagnosis secara otomatis dengan memanfaatkan data diagnostik pasien. Penelitian ini bertujuan untuk mengklasifikasikan risiko penyakit paru menggunakan berbagai algoritma machine learning pada aplikasi Orange3, serta menentukan model dengan akurasi terbaik. Dataset yang digunakan terdiri dari 5.000 data pasien dengan 18 atribut yang mencakup faktor demografis, gaya hidup, riwayat medis, dan kondisi klinis seperti kadar oksigen, tingkat stres, dan kebiasaan merokok. Lima algoritma diuji, yaitu iDecision Tree, Naïve Bayes, Support Vector Machine (SVM), k-Nearest Neighbor (kNN), dan Neural Network. Hasil pengujian menunjukkan bahwa Neural Network menghasilkan nilai akurasi tertinggi sebesar 89,15%, diikuti oleh Decision Tree (85,12%) dan Naïve Bayes (83,63%). Temuan ini membuktikan bahwa Neural Network lebih unggul dalam mengenali pola kompleks antarvariabel dan mampu memberikan prediksi yang lebih akurat. Dengan demikian, penelitian ini menegaskan potensi penerapan machine learning berbasis data diagnostik non-citra sebagai sistem pendukung keputusan untuk diagnosis dini penyakit paru.=================================================Abstract : Lung disease is a major contributing factor to high mortality rates worldwide. This is because lung disease is often difficult to detect in its early stages due to its nonspecific symptoms. The development of machine learning technology provides an opportunity to assist the automated diagnosis process by utilizing patient diagnostic data. This study aims to classify the risk of lung disease using various machine learning algorithms in the Orange3 application, and determine the model with the best accuracy. The dataset used consists of 5,000 patient data with 18 attributes covering demographic factors, lifestyle, medical history, and clinical conditions such as oxygen levels, stress levels, and smoking habits. Five algorithms were tested: Decision Tree, Naïve Bayes, Support Vector Machine (SVM), k-Nearest Neighbor (kNN), and Neural Network. The test results showed that Neural Network produced the highest accuracy value of 89.15%, followed by Decision Tree (85.12%) and Naïve Bayes (83.63%). These findings prove that Neural Network is superior in recognizing complex patterns between variables and is able to provide more accurate predictions. Thus, this study confirms the potential of applying machine learning based on non-image diagnostic data as a decision support system for early diagnosis of lung disease.

Copyrights © 2025






Journal Info

Abbrev

ijai

Publisher

Subject

Computer Science & IT

Description

Indonesian Journal of Applied Informatics publishes articles that are of significance in their respective fields whilst also contributing to the discipline of informatics as a whole and its application. Every incoming manuscript will first be examined by the Editorial Board in accordance with ...