Font Size: a A A

An Ontology-Based System For Projects And Domain Experts Matching

Posted on:2006-04-07Degree:MasterType:Thesis
Country:ChinaCandidate:G F YangFull Text:PDF
GTID:2156360152485514Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
It is very important to select the right experts for a certain project, and nowadays this kind of work is usually done by hand with oversimplified assistance of retrieval systems. In fact, these retrieval systems can only support keyword-based search, without understanding the meaning of the resource documents and the queries.In this paper, we introduce our ontology-based method to find the right expert who is matched well to a certain project. The word ontology has gained a great popularity within the information retrieval community, since ontology can improve the results quality comparing with the keyword-based method. We first formalize the documents which describing the projects and domain experts, with concepts and relations from the ontology. The formalization results are some concepts trees, and with an integrated and improved method we can calculate the semantic similarity between these concepts trees.The most significant works in this paper are:1. Integrating the Node-Based Approach and Edge-Based Approach, and giving an improved approach to calculating the semantic similarity between two nodes in two concept trees and between two concept trees.2. Developing a prototype system to compute the semantic similarity between project documents and expert documents, and sorting the matching results by the value of similarity.We develop our ontology draft with the classification of subjects, and formalize the documents searched from the Internet, with concepts and relations from mis ontology. We also design several tests to find whether our system can give promising outputs. The results of the tests show that our method can get better recall and precision. This kind of method presents a novel idea for matching projects and domain experts.
Keywords/Search Tags:Semantic Similarity, Semantic Matching, Ontology, Information Retrieval, 11-Setp Method
PDF Full Text Request
Related items