SABIO: Soft Agent for Extended Information Retrieval

Abstract

In the current study, an integrated system called SABIO is presented. The current system applies Information Retrieval (IR) techniques developed for collections of textual documents to nontextual corpa. SABIO integrates a fuzzy logic-based procedure for IR. Its search algorithm improves the IR efficiency and decreases the computational burden by using a fuzzy logic-based procedure for IR. This procedure is integrated in a flexible and fault-tolerant, human-reasoning-based search algorithm. The Accumulated Knowledge Set (AKS) of the system is sorted in a hierarchic multilevel tree-structure-like ontology. The objects in the AKS are represented using a novel human-reasoning-based-method. This representation takes into account the occurrence of related terms. The system uses a novel fuzzy logic-based term-weighting (TW) method. The developed fuzzy logic method improves the classical term frequency–inverse document frequency (TF=IDF) method, generally used for TW. The abovementioned system is the core of a wizard for search into the website of the University of Seville, www.us.es, which is currently in testing.