Anatomy and Empirical Evaluation of an Adaptive Web-Based Information Filtering System

Abstract

A case study in adaptive information filtering systems for the Web is presented. The described system comprises two main modules, named HUMOS and WIFS. HUMOS is a user modeling system based on stereotypes. It builds and maintains long term models of individual Internet users, representing their information needs. The user model is structured as a frame containing informative words, enhanced with semantic networks. The proposed machine learning approach for the user modeling process is based on the use of an artificial neural network for stereotype assignments. WIFS is a content-based information filtering module, capable of selecting html/text documents on computer science collected from the Web according to the interests of the user. It has been created for the very purpose of the structure of the user model utilized by HUMOS. Currently, this system acts as an adaptive interface to the Web search engine ALTA VISTA™. An empirical evaluation of the system has been made in experimental settings. The experiments focused on the evaluation, by means of a non-parametric statistics approach, of the added value in terms of system performance given by the user modeling component; it also focused on the evaluation of the usability and user acceptance of the system. The results of the experiments are satisfactory and support the choice of a user model-based approach to information filtering on the Web

Self-tuning Personalized Information Retrieval in an Ontology-Based Framework

“…measured in terms of precision and relevance) as subjectively perceived by users [7]. The key aspects involved include the representation of user interests (beyond a specific oneshot query), the dynamic acquisition of such interests by the system, and the exploitation of user preferences.…”

“…Personalized retrieval widens the notion of information need to comprise implicit user needs, not directly conveyed by the user in terms of explicit information requests [7]. Again, this involves modeling and capturing such user interests, and relating them to content semantics in order to predict the relevance of content objects, considering not only a specific user request but the overall needs of the user.…”

Personalized Search on the World Wide Web

“…The ordering of the first 30 results was considered. It shows that the system provides roughly a 34% improvement when compared to the search engine's non-personalized results (see [54] for details).…”

“…Many systems implement this approach on the client-side, e.g., [62,54,77], where the software connects to a search engine, retrieving query results that are then analyzed locally. In order to avoid spending time downloading each document that appears in the result list, the analysis is usually only applied to the top ranked resources in the list, or it considers only the snippets associated with each result returned by the search engine.…”

“…The Wifs system described in [54] is capable of filtering HTML or text documents retrieved by the search engine ALTAVISTA 3 in response to a query input by the user. This system evaluates and reorders page links returned by the search engine, taking into account the user model of the user who typed in the query.…”