A Retrieval System of Medicine Molecules Based on Graph Similarity

Abstract

Medicine information retrieval has grown significantly and is based on the structural similarity of medicine molecules. The chemical structural formula (CSF) is a primary search target as a unique identifier for each compound in the research field of medical information. This paper introduces a graph-based CSF retrieval system, PharmKi, which accepts the photos taken from smartphones and the sketches drawn on the tablet PCs as inputs. To establish a compact yet efficient hypergraph representation for molecules, we propose a graph-isomorphism-based algorithm for evaluating the spatial similarity between graphical CSFs. An indexing strategy based on the graph TF-IDF technology is also introduced to achieve a high efficiency for large-scale molecule retrieval. The results of comparative study demonstrate that the proposed method outperforms the existing methods on accuracy, and performs well on efficiency.