THIS PAPER PRESENTS A NEW EVALUATION APPROACH FOR AONTOLOGY-BASED EXPANSION STRATEGY PROPOSED IN PREVIOUS WORK. IN THIS WORK, THE EVALUATION FOCUSES ON THE SEARCH FOR LEARNING OBJECT (LO) IN SPECIALIZED REPOSITORIES. A TOTAL OF 98 USER QUERIES ONTHE GENE ONTOLOGY, A KNOWLEDGE REPRESENTATION LARGELY VALIDATED BY THE SCIENTIFIC COMMUNITY, WERE USED IN THE TEST. FOR EACH ONE, NEW CONCEPTS WERE EXTRACTED FROM THE ONTOLOGY CONSIDERING DIFFERENT EXPANSION TYPES. THE WORK CARRIED OUT DEMONSTRATES THAT QUERY EXPANSION DELIVERS BETTER RESULTS IN TERMS OF NOVELTY AND PRECISION, INDEPENDENTLY OF THE EXPANSION TYPE. ALSO, THAT EXPANSION TYPES RESULTING FROM ?PART-OF? RELATIONSHIPS PRESENT BETTER RESULTS IN TERMS OF PRECISION AND GAINED NOVELTY, IF COMPARED WITH EXPANSIONS USING ?IS_A? RELATIONSHIPS.
