Applying Expert Retrieval Model for Finding Expert Authors



This research applied expertise retrieval model for finding expert authors, and used evaluation methods of information retrieval systems for measuring the performance of those models. Current research is an experimental one. Besides, a variety of methods including survey method has been used in the research process. Various models were developed for finding expert authors, all built on a known expertise retrieval model, called document-centered model. Voting models were used within these models, too. All models were implemented and evaluated on the test collection developed specially for this research. Results showed that citation-centered models outperformed authorship-centered models in term of expert author finding task. In addition, there was no significant difference among retrieval performance of voting models. Individuals name occur in academic papers in two ways: either as author of paper or cited in the paper. Measuring retrieval performance of ER model considering kind of occurrence is our contribution in this research.