Skip to Main Content
The importance of patents is increasing as intellectual property becomes a core part of any industry. Most commercial patent retrieval systems are based on Boolean models, which are not capable of ranking similarity between queries and documents. In this study, we developed a patent retrieval system (PISS: patent information service system) which uses a vector space model and ontology to improve precision of search results and rank them by similarity. In other to evaluate the performance of query expansion, we tested ten queries, and found that the precision of search results was improved by an average of 36.2%. Ontology and additional search terms are automatically displayed next to search results for users' convenience.