An effective data indexing method for POI data

Abstract

The effective organization and management of POI data is an important basis for map service. In order to complete the auto-completion function, when users input some keywords part of an POI names to get some full name response, to improve the searching efficiency of Chinese POI data, we take the Chinese POI names as strings which are of high similarity and designed an effective data indexing method for Chinese POI names on the basis of a data structure--radix tree. So this work mainly deals with the approximate string search in large amounts of Chinese POI name strings. We present how to preprocess the Chinese POI name strings and store them in the radix tree. We then give an example to describe the procedure of input keyword string query based on the constructed radix tree of some Chinese POI names. In order to verify the effectiveness of the proposed method, we also analyzed the other two string query algorithms which are based on trie tree and ternary search tree respectively. We compare the performances between them from the perspective of time complexity and space complexity. Our results show that: 1) the performance of the POI search algorithm based on the radix tree is better than the other two algorithms both in the time complexity and space complexity for the Chinese POI names data; 2) the performance stability of the POI data search algorithm based on the radix tree is better than the latter two. The use of this kind of index mechanism makes the efficiency of Chinese POI data query respond well in Map Service which will have a promising future.