A natural language processing based Internet agent

Searching for useful information is a difficult job by the virtue of the information overloading problem. With technological advances, notably the World-Wide Web (WWW), it allows every ordinary information owner to offer information online for others to access and retrieve. However, it also makes up a global information system that is extremely large-scale, diverse and dynamic. Internet agents and Internet search engines have been used to deal with such problems. But the search results are usually not quite relevant to what a user wants since most of them use simple keyword matching. In this paper, we propose a natural language processing based agent (NIAGENT) that understands a user's natural query. NIAGENT not only cooperates with a meta Internet search engine in order to increase recall of web pages but also analyzes the contents of the referenced documents to increase precision. Moreover, the proposed agent is autonomous, light-weight, and multithreaded. The architectural design also represents an interesting application of a distributed and cooperative computing paradigm. A prototype of NIAGENT, implemented in Java, shows its promise to find more useful information than keyword based searching.