The Efficient Geospatial Index Structure for Spatial Keyword Query with Myanmar Language on Mobile Devices

Abstract

Nowadays, Location-based services (LBS) are widely used in numerous areas such as positioning systems, tourist information, social and business areas. Many researchers try to develop the LBS application to fulfill the user requirements. The usage of location-based services systems is increasing around the world including Myanmar. Tourists in Myanmar are now increasing day by day. Location-based services systems in Myanmar still have the requirements compared with the other developed countries to provide both local users and foreign users with the geo-information of Myanmar like nearest geo-information, geo-information around the user’s desired point. Moreover, most of the location-based services applications are processed in online. Sometimes, it cannot access the information when the internet connection is weak or cannot be available. This research is implemented for retrieving the geo-information on the mobile devices using both Myanmar and English language to provide both local and foreign visitors although it is considered for mainly focus on the Myanmar language. Therefore, it can support not only foreign users but also local users who are difficult for using the English version of the location-based services applications. For fetching the geo-information, the nearest keyword query and range keyword query can be operated not only from the user’s current location but also from the user’s desired place. These queries are processed based on the given location and given keyword. In this research, user can pre-find the desired geo-location due to be able to process these queries from the user’s desired location. In addition, the geo-information can be obtained by using the keyword search. In this keyword query, the geo-information is not only the location information with their related text information but also the distance information from this geo-information to the user’s current location. Moreover, it also gives the geo-information which has the minimum distance to current location among the resulted geo-information. These spatial keyword queries can be processed whether the internet connection is available or not. The geo-information contains not only the geo-location (latitude and longitude) but also the related text information. The data of the geo-information can be huge so that it can be time-consuming for searching the geo-information. The geo-spatial index structure is one of the key factors to perform the fast spatial keyword query processing because the mobile devices possess the limited memory and a low computationaliii capacity than the personal computer. Thus, it is required to build the geospatial index structure to obtain the efficient spatial keyword query processing. In this research, a new index structure is constructed to get the fast access the geo-data for Myanmar keyword query. In addition, this index structure is also considered for extracting the geo-information with English keyword Query. Hilbert curve, B-tree and inverted file are utilized for constructing this index structure to get the efficient searching on both geo-location data and textual data. Due to B-tree works on the one-dimensional data, Hilbert curve is used to map the two-dimensional data (geo-location data) to one-dimensional data. Then, B-tree is built according to the Hilbert value and inverted file is embedded to the B-tree node. The proposed index structures are separately constructed depend on the services such as school, hotel and so on. Geo-spatial keyword queries processing in this research are operated with this proposed index structure. It can reduce the searching time by using this proposed index structure. This research focuses on the Yangon Region. Mapbox vector map is used to display the geo-information in online and offline. Myanmar 3 font is used in this research. This location-based service application can be used in android mobile version 4.4 and above.