Application of machine learning techniques for analysing molecular emission lines

Modern astronomy deals with enormous set of spectral data obtained from observations of molecular emissions in various sources—including infrared dark clouds, protostars, and HII regions. Traditional data analysis methods become less effective when processing such extensive datasets. In this work, we explore the application of machine learning techniques for analysing molecular emission lines to characterise astronomical objects and their astrochemical properties. Utilising data from the MALT90 survey, which observes the Galactic plane at 3 mm wavelengths, we discuss various clustering algorithms and dimensionality reduction methods that simplify the identification of objects. Our findings demonstrate that machine learning methods effectively identify two groups of star-forming regions among other types of astronomical sources. These groups are distinguished by line intensities of molecules which have different astrochemical origins. This approach holds promise for processing data collected with advanced facilities like JWST and ALMA, leading to more efficient analysis of astronomical data.