Accurate and quantitative classification of gas mixtures is an important issue in various fields, including the healthcare and food industries. However, traditional classification approaches such as gas chromatography, mass spectroscopy, and chemical analysis not only require specialized skills but are also time-consuming, inaccurate, and expensive. For these reasons, we used a chemiresistive sensor based on 2D transition metal dichalcogenides and platinum group material based chalcogenides, which have high responsivity, selectivity, and stability toward target gases. Raman spectroscopy, scanning electron microscopy, and X-ray photoelectron spectroscopy were used to characterize the WS2 and RuS2 sensing channels. Moreover, the gas-sensing properties toward NO2, NH3, and their mixtures (1:1 and 2:1) were analyzed, and the classification of these gases was carried out via our proposed two-stage classification model consisting of dimensionality reduction and classification processes. The proposed model achieved more than 90 % accuracy in all cases when classifying single gases and their mixtures, which could be industrially applicable in the future. IEEE