Searching Information Servers Based on Customized Profiles

We investigate the effect of using customized profiles to help searching relevant servers in Internet. Our experiments demonstrate that the use of customized profiles with latent semantic indexing (LSI) technique can improve the performance of Internet searching. 1 Introduction When searching information in a retrieval system, people use different terms to describe their information needs. The retrieval system searches through its database and returns documents indexed with matching terms. Since a concept can be represented by a variety of terms, users may fail to obtain the information they require. This is called the vocabulary problem [1]. The vocabulary problem occurs not only in traditional information retrieval, but also in Internet resource discovery, where users seek relevant information servers to submit their queries. Previously, we proposed to use Latent Semantic Indexing (LSI) [2] to ameliorate the vocabulary problem in the Internet search [3]. Here, we expand the idea by ..