Machine Learning interpretation of the correlation between infrared emission features of interstellar polycyclic aromatic hydrocarbons

TL;DR

This paper uses machine learning to analyze and interpret the physical basis of infrared emission features of interstellar polycyclic aromatic hydrocarbons, revealing molecular fragments responsible for observed astronomical correlations.

Contribution

It introduces a feature importance analysis method with random forests to connect emission features to molecular structures, enhancing understanding of interstellar PAH spectra.

Findings

Identified molecular fragments linked to specific infrared bands.

Demonstrated the effectiveness of extended-connectivity fingerprints for prediction.

Mapped correlations between emission bands based on feature importance similarities.

Abstract

Supervised machine learning models are trained with various molecular descriptors to predict infrared emission spectra of interstellar polycyclic aromatic hydrocarbons. We demonstrate that a feature importance analysis based on the random forest algorithm can be utilized to explore the physical correlation between emission features. Astronomical correlations between infrared bands are analyzed as examples of demonstration by finding the common molecular fragments responsible for different bands, which improves the current understanding of the long-observed correlations. We propose a way to quantify the band correlation by measuring the similarity of the feature importance arrays of different bands, via which a correlation map is obtained for emissions in the out-of-plane bending region. Moreover, a comparison between the predictions using different combinations of descriptors underscores the strong prediction power of the extended-connectivity molecular fingerprint, and shows that the combinations of multiple descriptors of other types in general lead to improved predictivity.