A graphics processing unit-based parallel algorithm for keyword search on a data graph

Keyword search on a data graph has attracted increasing attention due to its practicability. With continuous expansion of the amount of data to be searched, common search methods have been unable to satisfy the demands of users for efficiency. Aiming at improved search efficiency, this study proposed a graphics processing unit (GPU)-based parallel algorithm for keyword search on data graphs. This method searched by establishing a keyword index and a path index. The keyword index therein was used to position keyword nodes, while the path index was used to search the results. Meanwhile, a CPU-based parallel algorithm was introduced in the establishment of index and keyword search. The experimental results suggested that it was effective at improving keyword search efficiency. Copyright © 2015 Binary Information Press.