Generating User Profiles by Translating Content Queries to Concepts Using Thesaurus

Abstract

As more information becomes available electronically, information retrieval or filtering tools for finding information of interest to users become increasingly important. Building tools for assisting users in finding relevant information is often complicated by the difficulty in articulating user interest in a form that can be used for searching. The aim of the approach described here is to build instantaneous user profiles in a single episode information-seeking environment that instantaneously and accurately captures uses interests with minimum user interaction. The research work described here focuses on the importance of accuracy and preciseness of user profiles that could be represented in terms of concepts, i.e., keywords or subject descriptors assigned to technical documents. Studies conducted here are in the context of understanding and building content-based profiling system for technical documents, specially in the field of aerospace science and technology. Sample data of aerospace grey literature in the form of technical papers with title and associated descriptors have been considered for the design and development of database and webbased profile builder for generating user profiles. The descriptors assigned to technical papers with the help of technical thesaurus were considered. Equal importance or weightage was given to broader and narrower terms, because there is no pre-defined mechanism to assign different weightage to different subject descriptors. Generated user profiles in terms of weighted vector of descriptors could directly be used for recommending technical documents with relevance ranking for latest addition of documents.