|
Knowledge organization (KO) is a tradition rooted in libraries where information is organized according to some knowledge schemes, such as taxonomies, thesauri, gazetteers, etc. It had been the dominant way for information discovery and utilization before the Web age. However, tremendous human efforts are demanded in the building, maintenance and employment of KO tools. Unable to scale to massive information situation, it loses the popularity in Web environment, thus making search engines become the dominant tool for information discovery and utilization, even the only one to most of Web roamers.
To provide richer and powerful Web application tools, and to enhance the power of search engines, the tradition of knowledge organization should be revivified in Web environment. This is why ontology and folksonomy draw more and more attention in research communities and industry world.
This talk introduces some pioneer works conducted by the KVision research group at the Dept. of Information Management of Peking University on reconstructing traditional KO tools for Web applications, including:
1) Automatic enriching thesauri with new terms extracted from metadata;
2) Automatic reconstructing large-scale classification schemes (e.g. Dewey Decimal Classification) for automatic classification;
3) integrating traditional KO resources to automatically construct an ontology, and implementing semantic retrieval and knowledge navigation with an ontology-driven approach;
4) employing the reconstructed KO tools to augment the power of search engines with vocabulary aiding and search result clustering.
At last, a Chinese social bookmarking tool Doulor, which is developed recently in the KVision lab, will be demonstrated. It implements the findings from previous research projects and combines the power of folksonomy with ontology, thus offering a Web2.0-style resolution for Web information organization.
|
Tel: +852 2609-8313