A Web Search Method Integrating Taxonomy-based and Crawler-based Search Engines
スポンサーリンク
概要
- 論文の詳細を見る
With the rapid advance of the Internet technology, efficient information retrieval in the web space has been an important research issue. One way to search the information needed from the web space is to use taxonomy. In this paper, we propose a dynamic web search method which integrates existing taxonomy-based search engines and crawler-based search engines. In the proposed scheme, the user can search for information stored in crawler-based search engines utilizing a taxonomy provided by another existing taxonomy-based search engine. First, the user gives a query and selects a context category on the taxonomy. The system then constructs a rule-based classifier using pre-classified pages in the taxonomy-based search engine. The classifier is constructed dynamically and on demand based on pages matching the query and the selected context category. Next, the system uses the classifier to modify the query in a simple manner. Finally, the modified query is sent to the crawler-based search engines and the results are presented to the user. Our approach has the following features: (1) It provides taxonomy-based search facility in a variety of context by making the best of existing taxonomy-based search engines. (2) By combining the two types of the search engines, we can increase the coverage of the existing taxonomy-based search engines. (3) Since our classifier is constructed dynamically, it can reflect the user intent and up-to-date taxonomy structures. To evaluate the effectiveness of our method, we conduct some experiments on the existing taxonomy-based and crawler-based search engines.
- 一般社団法人情報処理学会の論文
- 2002-09-15
著者
-
Pahlevi Said
Doctoral Program In Engineering University Of Tsukuba
-
Kitagawa Hiroyuki
Institute Of Information Science And Electronics University Of Tsukuba
関連論文
- False Drop Analysis of Set Retrieval with Signature Files
- Requirement Specification and Derivation of ECA Rules for Integrating Multiple Dissemination-Based Information Sources (the 2002 IEICE Excellent Paper Award)
- Requirement Specification and Derivation of ECA Rules for Integrating Multiple Dissemination-Based Information Sources
- Design and Performance Analysis of Indexing Schemes for Set Retrieval of Nested Objects
- Optimization of Join-Type Queries in Nested Relational Databases
- Join Query Optimization in Object-Oriented Database
- A Web Search Method Integrating Taxonomy-based and Crawler-based Search Engines