MSQ-Index: A Succinct Index for Fast Graph Similarity Search

Graph similarity search under the graph edit distance constraint has received considerable attention in many applications, such as bioinformatics, data mining, pattern recognition and social networks. Existing methods for this problem have limited scalability because of huge amount memory they consume when handling very large databases with tens millions graphs. In article, we present a succinct index that incorporates structures hybrid encoding to achieve improved query time performance minimal space usage. Specifically, usage our requires only 5-15 percent previous state-of-the-art indexing size while at same achieving several times acceleration on tested data. We also improve by augmenting global filter range searching, which allows us perform reduced region. addition, propose two effective lower bounds together boosting technique obtain smallest possible candidate set. Extensive experiments demonstrate proposed approach is superior both filtering approaches. To best knowledge, first in-memory successfully scales cope dataset 25 million chemical structure graphs from PubChem dataset. The source code available online.