A metric for pairwise similarity analysis of binary cheminformatics data
Abstract: This paper unveils the findings derived from an in-depth exploration of a novel similarity measure designed to assess pairwise resemblances. Called the Substructure Similarity Index, this measure centers around the comparison of substructures identified within compounds. Through a rigorous evaluation conducted on an extensive dataset of drugs and by juxtaposing it against other commonly employed indices, the study reveals that the Substructure Similarity Index can be adeptly employed for molecular similarity calculations since it provides information that cannot be obtained by available measures.
engleski
2023
© All rights reserved
Keywords: molecular similarity, molecular structure, binary vectors, molecular fingerprints, similarity measure.