| Specialization and intelligence has become a focus of attention, but search engine cannot offer specialized information services to customers any more. Question-Answer(referred as Q-A system)is a kind of efficient, flexible, humanized information retrieval systems.As a newly-developed pattern of knowledge sharing, Q-A system is prevalent for its interactivity and openness.Although Q-A system is well-developed, few researches of Q-A system on courses have hitherto been conducted. Applying Q-A system to teaching can not only increase the efficiency of study and knowledge sharing, but also raise the study interests of students, which has a positive effect on educational quality. In this paper, research is conducted on a curriculum Q-A system based on community Q-A technology. Such a system enables users to describe questions by natural language, subsequently return the corresponding answer in brief and precision. Under the circumstance of no suitable answers, this system will transfer the question to relevant experts for their answers.Firstly, initialize the Database of curricular Q-A system. The initialization aims to fetch the existing answers of curricular knowledge from the Internet, which will be stored in the database after text processing, waiting for information retrieval.Secondly, analyze questions. This article uses the ICTCLAS, developed by Chinese Academy of Sciences Institute of Computing Technology, to complete the word segmentation and part-of-speech tagging and subject extraction. Additionally, synonyms method is used for subject extension, also combination of Cosine similarity based on spatial vector model and extension of synonymous dictionary is used for classification of interrogative sentences.Thirdly, retrieve similarity questions. This paper uses method of Similarity Calculation to find out the questions which are similar or same to the question. Referring to the existing Similarity Calculation method, Similarity Calculation method based on key words match was put forward. Rank the candidate problem sets of information retrieval on the basis of the similarity degree from big to small sort. The method takes the surface information of the question into account, and combined with some semantic information.Fourthly, recommend expert users. When user searches the similar questions which system does not find, it stores this question in the Database of curricular Q-A system and then transfers this kind of questions to relevant expert users for answers. The system will give back this answer to user.Finally, the prototype of the CQA system based on Curriculum Knowledge is designed and implemented. According to the test result, it runs well. |
