Efficient Combined Index Structure for K-Nearest Neighbours Keyword Search on Spatial Database

Abstract

Spatial keyword search on spatial database has been well studied for years due to its importance to commercial search engines.Specially, a spatial keyword query takes a user location and user-supplied keywords as arguments and returns object that is nearest k objects from user current location and textually relevant to the user required keyword. Geotextual index play an important role in spatial keyword querying. This paper proposes the efficient combined index structure for K-Nearest Neighbours Keyword Search on Spatial Database. That combine K-d tree and inverted file for nearest neighbor keyword query which is based on the most spatial and textual relevance to query point and required keyword. It can search required k results with minimum IO costs and CPU costs. The k-results are ranked according to the distance or keyword. The own dataset is created for Yangon (Myanmar) region which contains latitude, longitude, name, description and category type of each object.