Electronic Theses and Dissertation

Penyakit Paru Obstruktif Kronik (PPOK) adalah kondisi paru-paru jangka panjang yang memerlukan deteksi dini tanpa menggunakan metode invasif. Salah satu cara untuk mendeteksinya secara dini adalah dengan menganalisis senyawa kimia dalam napas seseorang yang disebut Volatile Organic Compounds (VOCs). Penelitian ini mengembangkan sistem untuk mengklasifikasikan PPOK menggunakan machine learning dan data VOCs dari sampel napas. Data dibagi menjadi 80% untuk melatih model dan 20% untuk pengujian. Berbagai metode machine learning digunakan, termasuk Logistic Regression, K-Nearest Neighbor (KNN), Support Vector Machine (SVM), dan Random Forest. Model-model tersebut diuji menggunakan beberapa metrik seperti accuracy, precision, recall, F1-score, confusion matrix, ROC Curve, dan nilai AUC. Hasil menunjukkan bahwa model Random Forest bekerja paling baik, dengan acuracy 86% danF1-score rata-rata 86%, serta nilai AUC 0,94. Model ini mampu membedakan jenis-jenis PPOK berdasarkan data VOCs napas. Hasil ini menunjukkan bahwa penggunaan machine learning dengan analisis VOCs dari napas dapat menjadi alat yang berguna untuk deteksi PPOK secara noninvasif.

Kata Kunci : PPOK, VOCs, e-Nose, Logistic Regression, K-Nearest Neighbor, Random Forest, Support Vector Machine, Machine Learning

Chronic Obstructive Pulmonary Disease (COPD) is a serious lung condition that lasts a long time. It's important to find it early without using painful or harmful tests. One method to find COPD early is by looking at certain chemicals in a person's breath known as Volatile Organic Compounds (VOCs). This study created a system that uses machine learning and VOC data from breath samples to classify COPD. The data was split into two groups: 80% was used to train the system, and 20% was used to test how well it worked. Different machine learning techniques were tested, such as Logistic Regression, K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Random Forest. The performance of each model was measured using several tools like accuracy, precision, recall, F1-score, confusion matrix, ROC Curve, and AUC value. The results showed that the Random Forest model did the best job. It had an accuracy of 86%, an average F1-score of 86%, and an AUC value of 0.94. This model could also tell the difference between different types of COPD based on the VOCs in the breath. These findings suggest that using machine learning with breath VOC analysis can be a helpful tool for detecting COPD without invasive methods. Keywords: COPD, VOCs, e-Nose, Logistic Regression, K-Nearest Neighbor, Random Forest, Support Vector Machine, Machine Learning.