The Development of Ontology-based Intelligent Retrieval Algorithm and Prototype System of Mechanical Products

For the better and stronger information retrieval function, the sequence of search result queue and the correlation of query expression are the key problems that could be computed by similarity algorithm. A new method which is ontology-based weighted semantic tree similarity algorithm (OWSTS) is presented here. In this paper, the construction method of OWSTS and the implementation of similarity algorithm are introduced firstly. And an ontology-based information retrieval prototype system of mechanical products domain is developed secondly. Finally the relative accuracy of retrieval results by OWSTS and generally used cosine vector space based similarity algorithm (TF*IDF) are compared. The successful running of this system shows that OWSTS algorithm is feasible for information retrieval and the similarity algorithm makes the sequence of search result queue fit for users' semantic requirement.