In this paper, we propose a syntax-based text retrieval system for on-line language learning and use a fast regular expression search engine as its main component. Regular expression searches provide more scalable querying and search results than keyword-based searches. However, without a well-designed index scheme, the execution time of regular expression search would be unacceptable to users. Our methods are based on Cho and Rajagopalan (2002) and we introduce some modifications, such as a presuf index constructing algorithm and a method for deciding minimum filter factor, to meet the requirements of our syntax-based text retrieval system. The experiment results show the index space size is small and the performance of syntax-based sample queries show significant improvements over benchmark results. A user-friendly query generator is designed to support users who have no background knowledge of regular expressions.
Educational Technology and Society 12(1), pp.73-81