Electronic Nose-Based Classification of Breath Odor in Fasting and Non-Fasting Individuals Using Principal Component Analysis

Imam Tazi (Universitas Islam Negeri Maulana Malik Ibrahim Malang)
Muthmainnah (Universitas Islam Negeri Maulana Malik Ibrahim Malang)
Wiwis (Universitas Islam Negeri Maulana Malik Ibrahim Malang)
Yahya (Universitas Islam Negeri Maulana Malik Ibrahim Malang)
Yusril (Universitas Islam Negeri Maulana Malik Ibrahim Malang)

Jurnal Penelitian Pendidikan IPA (JPPIPA)
Vol 12 No 6 (2026)

Breath odor contains volatile compounds that can reflect the body’s metabolic condition, including those associated with fasting. This study aims to classify breath odor patterns in fasting and non-fasting individuals using an electronic nose (e-nose) combined with the Principal Component Analysis (PCA) method. The e-nose system was developed using an array of five MQ gas sensors (MQ-2, MQ-3, MQ-4, MQ-5, and MQ-6) to detect breath samples from 100 respondents, consisting of 50 fasting and 50 non-fasting individuals. The sensor responses were recorded and analyzed using PCA for dimensionality reduction and data pattern visualization. The results show that the first two principal components account for 81.2% of the total data variance, with contributions of 56.1% from PC1 and 25.1% from PC2. The PCA score plot demonstrates a relatively clear separation between the breath odor patterns of the fasting and non-fasting groups. These findings indicate that the developed e-nose system has potential as a rapid and non-invasive method for breath-based classification.