C.-L. Huang and C.-C. Chiang (Taiwan)
Electronic document management, information retrieval, ISO documents, vector space model, personalization
This study proposes a personalized document query system. The personalized ranking rules which incorporate the user profile and document profile were combined with the traditional information retrieval method to improve the query quality. The queried documents were obtained using a traditional information retrieval method, and re-ranked by the personalized ranking rules. This approach was applied to the ISO document retrieval system. A prototype ISO document retrieval system was implemented based on Lotus Notes in a high-tech electronic components manufacturing company. The prototype system provided three types of query approaches: full text keyword search, a pure vector space model (VSM), and the proposed personalized model—VSM with personalization. An empirical experiment showed that the VSM with personalization model achieved better performances than the pure VSM and full text keyword search under various sizes of top-N recommendations.
Important Links:
Go Back
