Automated Mixture Analysis via Structural Evaluation

• DOI: 10.48550/arxiv.2408.15819
• Publication Date: 2024-08-28
• Authors: Zachary T. P. Fried; Brett A. McGuire
• Keywords: FOS: Computer and information sciences; Computer Science - Machine Learning; Astrophysics of Galaxies (astro-ph.GA); FOS: Physical sciences; Astrophysics - Astrophysics of Galaxies; Machine Learning (cs.LG)

The determination of chemical mixture components is vital to a multitude scientific fields. Oftentimes spectroscopic methods are employed decipher the composition these mixtures. However, sheer density spectral features present in databases can make unambiguous assignment individual species challenging. Yet, commonly chemically related due environmental processes or shared precursor molecules. Therefore, analysis relevance molecule important when determining which mixture. In this paper, we combine machine-learning molecular embedding with graph-based ranking system determine likelihood being based on other known and/or priors. By incorporating metric rotational spectroscopy algorithm, demonstrate that be identified extremely high accuracy (>97%) an efficient manner.