Relevance is the key in defining what data is to be extracted from the Internet. Traditionally, relevance has been defined mainly by keywords and user profiles. In this paper we discuss a fairly untouched dimension to relevance: location. Any navigational information sought by a user at large on earth is evidently governed by his location. We believe that task oriented data mining of the web amalgamated with location information is the key to providing relevant information for personal navigation. We explore the existential hurdles and propose novel approaches to tackle them. We also present naive, task-oriented data mining based approaches and their implementations in Java, to extract location based information. Ad-hoc pairing of data with coordinates (x, y) is very rare on the web. But if the same co-ordinates are converted to a logical address (state/city/street), a wide spectrum of location-based information base opens up. Hence, given the coordinates (x, y) on the earth, the scheme points to the logical address of the user. Location based information could either be picked up from fixed and known service providers (e.g. Yellow Pages) or from any arbitrary website on the Web. Once the web servers providing information relevant to the logical address are located, task oriented data mining is performed over these sites keeping in mind what information is interesting to the contemporary user. After all this, a simple data stream is provided to the user with information scaled to his convenience. The scheme has been implemented for cities of Finland.

Date Published: 12 March 2002
PDF: 11 pages
Proc. SPIE 4730, Data Mining and Knowledge Discovery: Theory, Tools, and Technology IV, (12 March 2002); doi: 10.1117/12.460246

Published in SPIE Proceedings Vol. 4730:
Data Mining and Knowledge Discovery: Theory, Tools, and Technology IV
Belur V. Dasarathy, Editor(s)