Proposta de modelo de colaboração para catálogo web facetado

Abstract

This study proposes a collaborative model for indexing and searching for items in a collaborative faceted web catalog (Facetlog), where its own users create the content. It aims at enhancing or improving the quality of indexing by offering a complementary strategy to the tagging through the selection of terms from a faceted taxonomy. It also aims at perfecting the information retrieval in web catalogs by using a searching interface that combines faceted and keyword search. In order to design the collaborative model, a review of the literature was conducted on the following topics: a) processes and models of information retrieval; b) indexing through controlled vocabulary; c) faceted classification; d) faceted taxonomy; e) automatic indexing; f) faceted searching interface; g) faceted web catalogs; h) Web 2.0 collaboration models; i) collaborative systems; j) workflow; k) moderation; l) social tagging; m) wiki; and n) voting systems. Studies on faceted tagging, collaborative faceted classification and faceted yellow pages were also researched in the review of the literature. The collaborative model was created by using the RUP software development process, the UML language diagram, low-fidelity prototyping and open source softwares like the WordPress. The evaluation of the collaborative model, on the other hand, was performed the usability test to validate the requirements. This test revealed the userss preference to use the keyword search instead of the faceted search. It also showed their inclination to use tagging instead of faceted taxonomy. However, the study also revealed that both the faceted search and the faceted taxonomy are valid complementary strategies to information retrieval and indexing, given that both are used in one-third of the userss actions. This study concludes by recommending the use of the autocomplete function for selecting tags and categories in the searching and indexing interfaces. It also recommends that the level of moderation should vary according to the degree of impact of an action on the faceted web catalog. The study suggests the following pattern: voting, comment and classification shoud be associated to a pre-approved moderation and the taxonomy management to an approved moderation.