TY - GEN
T1 - Agents for user-profiling, information filtering, and information monitoring
AU - Sim, Kwang Mong
AU - Kwok, Paul C.K.
N1 - Funding Information:
This project is designed to support the information gathering activities in Ellis’ model [] that are not supported by existing search engines. Ellis’ model [] of information gathering includes (1) activities that form initial search for information by following and linking to other information sources, (2) browsing (scanning information source), (3) differentiating (filtering and selecting among the sources), (4) monitoring (regularly following a particular source), and (5) extracting (identifying materials of interest from some sources). Whereas activity (1) is supported by existing search engines, the enhanced holistic IR system in this project is designed to bolster activities (2) through (5). To this end, this project complements and augments the functionalities of existing search engines. In particular, it is reminded here that this project does not compete with existing search engines and is certainly not designed to replace existing search engines, but rather to supplement their functionalities.
PY - 2010
Y1 - 2010
N2 - This paper presents an enhanced holistic information retrieval (IR) system that aims to automate the entire process of Web-based IR. The system consists of four types of agents: (1) a User profiling agent (UPA) that filters and reorders URLs based on a user's interests, (2) ontology-enhanced Web browsing agents (WBAs) that are used to autonomously browse and scan multiple Websites to determine and rate the relevance of Websites, (3) Web monitoring agents (WMAs) that are used for tracking and reporting changes in selected Websites, and (4) price watcher agents (PWAs) that monitor product prices from competing suppliers' Websites. A UPA generates a profile of a user's interests, then filters and reorders URLs based on the interests of the user. WBAs perform information filtering by considering three relevance metrics: ontological relations, frequency, and nearness of keywords. The general idea of Website monitoring is that each WMA is programmed to download a new copy of a Website and compare it with the old copy. WMAs allow users to specify monitoring rules, and provide user interface for specifying patterns and data to be monitored. PWAs invoke the functionalities of WBAs and WMAs for browsing and monitoring multiple Websites displaying different prices of a product. Whereas empirical results show that WBAs are likely to rate the relevance of Website with a small degree of error, the UPA can generally identify URLs that a user is more likely to be interested in. Proof-of-concept examples demonstrate the major functionalities of WMAs and PWAs.
AB - This paper presents an enhanced holistic information retrieval (IR) system that aims to automate the entire process of Web-based IR. The system consists of four types of agents: (1) a User profiling agent (UPA) that filters and reorders URLs based on a user's interests, (2) ontology-enhanced Web browsing agents (WBAs) that are used to autonomously browse and scan multiple Websites to determine and rate the relevance of Websites, (3) Web monitoring agents (WMAs) that are used for tracking and reporting changes in selected Websites, and (4) price watcher agents (PWAs) that monitor product prices from competing suppliers' Websites. A UPA generates a profile of a user's interests, then filters and reorders URLs based on the interests of the user. WBAs perform information filtering by considering three relevance metrics: ontological relations, frequency, and nearness of keywords. The general idea of Website monitoring is that each WMA is programmed to download a new copy of a Website and compare it with the old copy. WMAs allow users to specify monitoring rules, and provide user interface for specifying patterns and data to be monitored. PWAs invoke the functionalities of WBAs and WMAs for browsing and monitoring multiple Websites displaying different prices of a product. Whereas empirical results show that WBAs are likely to rate the relevance of Website with a small degree of error, the UPA can generally identify URLs that a user is more likely to be interested in. Proof-of-concept examples demonstrate the major functionalities of WMAs and PWAs.
KW - Software agent
KW - Web information retrieval
UR - http://www.scopus.com/inward/record.url?scp=84867286573&partnerID=8YFLogxK
U2 - 10.1007/978-90-481-3517-2-35
DO - 10.1007/978-90-481-3517-2-35
M3 - Conference contribution
AN - SCOPUS:84867286573
SN - 9789048135165
T3 - Lecture Notes in Electrical Engineering
SP - 457
EP - 474
BT - Intelligent Automation and Computer Engineering
T2 - International Conference in Intelligent Automation and Computer Engineering, Under the Auspices of the International MultiConference of Engineers and Computer Scientists, IMECS 2009
Y2 - 18 March 2009 through 20 March 2009
ER -