| Summary: | 碩士 === 國立高雄第一科技大學 === 資訊管理研究所 === 104 === The writing ability is one of the necessary skills in every field. But we always feel confused when we start writing. Usually, we can not express clearly what it means in the upcoming sentences, caused we do not used to write scientific papers. So we developed a system with automatic candidate words selection and example sentences recommendation system, which can improve our writing performance.
In this paper, we use “Google Web 1T N-Gram Version 1” and Google search keywords to setup our data corpus. Elasticsearch is used as our search engine and the Memcached is used to make the searching mechanisms faster. Finally, we have built a paper editor with Automatic Candidate Words Selection function and sample sentences recommendation system.
When the system is finished, we conducted a user satisfaction survey and Elasticsearch cluster benchmarking. We use 2, 4, 8 and 16 hosts to test response time. For satisfaction surveys, users for the system overall ratings have nearly 75% of users rated satisfactory. It proved that the proposed system can efficiently improve the paper writing speed. For Elasticsearch benchmarking, we using “must” and “must_not” search script to analyzing response time on different hosts. In 16 hosts case, the response time is 0.012s, less than 8 hosts (0.16s), and less than 4 hosts (0.237s).
|
|---|
