Query expansion and layers filtration in tourist spatial information retrieval / Rosilawati Zainol

Increasing demand in retrieving information with geographic reference has challenged scholars to conduct studies in producing efficient method in retrievingspatial data and evaluating web-based geographic information system (GIS) on the internet, also known as web-based GIS. Tourist relies on maps throughout his visits in order to fulfil his information needs. Online maps with geographic information system (GIS) capabilities that are able to cater for tourist's spatial needs are few on the internet. Therefore, this research's objectives are to examine online map preferences by Malaysians, to evaluate the current online GIS maps capabilities, and to discover an efficient method to retrieve spatial information required by tourist. In order to achieve these objectives, tourist's spatial information databases of Shah Alam City Centre, Kuala Lumpur Golden Triangle and Taiping Historical Town Centre are developed. Furthermore, since spatial data are stored in separate layers according to their features: point, polyline and polygon, gazetteer method is found to be the most suitable method of handling query of this data type. Therefore, exact match and partial match keyword queries are used. Query lists are collected from surveys. A survey on online map preference is carried out among 489 Malaysians in Klang Valley to determine their preferences in conducting online search on spatialinformation. This survey contributes to the collection of query lists and relevant judgment. In addition, relevant judgment is also derived from interviews. Four experiments were carried out using all test collections. First experiment examines query using GoogleMaps Malaysia. The second experiment examines query by using the original Super WebGIS 2.1. The third experiment examines query using enhanced Super WebGIS 2.1 by case insensitivity and layer filtration. Finally, the fourth experiment examines query using enhanced Super WebGIS 2.1 by query expansion and the combination of the third technique. Results show that combination of query expansion, case insensitivity and layer filtration technique yields 70% better results than using case insensitivity and layer filtration alone and 98% better than without using any of the three. The combination of these experiments has produced a system named MuSafIR (Musafir Spatial Information Retrieval).