自学考试网上学习社区中交互式答疑系统的设计和实现

发布时间：2019-01-04 09:07

【摘要】：交互式答疑系统的构建是一个复杂系统的工程，，包含的功能众多，研究的侧重点主要在于如何提高机器答疑的智能性及机器答疑与人工答疑的无缝链接。提高机器答疑智能性的一个重要方面是使系统能够准确理解学习者提出的问题，为此，一方面通过改进中文分词器提高系统对中文分词的能力，另一方面，通过构建学科概念本体库，提高系统理解中文语义的能力。机器答疑与人工答疑的无缝链接通过后台统计记录及衔接功能，将机器答疑中解决不了的问题自动转发给人工答疑模块，由人工答疑模块根据问题类型选择相应的教师实施人工答疑。系统会将人工答疑的结果会反馈给机器答疑模块，以不断提高机器答疑模块答案库的容量。根据答疑系统的功能要求及自考生的实际检索需求，对全文搜索引擎Lucene自带的中文分词器进行了改写。在为构建资源索引时，系统同时使用学科本体库与Lucene自带的词库对词语进行分词，进而为分割后的词语建立索引。在检索资源时，答疑系统会首先使用本体库中的概念对用户输入的内容进行分词，进而根据这些分割后的关键词到系统的各个知识库索引中查找相关内容，然后，再使用Lucene自带的分词器对用户的输入进行分割，使用分割后的关键词再去索引库中查找相关内容，最后，检索系统将两次检索的结果合并，使用学科本体库概念查找的结果排在前面，使用由Luence自带分词器分割的关键词查找的资源排在后面呈现给检索者。
[Abstract]:The construction of interactive question answering system is a complex system with many functions. The focus of the research is how to improve the intelligence of machine answering and the seamless link between machine answering and manual answering. An important aspect of improving the intelligence of machine answering questions is to enable the system to accurately understand the problems raised by learners. For this reason, on the one hand, the ability of the system to Chinese word segmentation can be improved by improving the Chinese word segmentation device, on the other hand, By constructing the subject concept ontology database, we can improve the ability of understanding Chinese semantics. The seamless link between machine answering and manual answering can automatically forward the problems that can not be solved in machine answering to the manual answering module through the backstage statistical record and the function of connecting with each other. According to the type of question, the manual answering module selects the corresponding teachers to answer questions manually. The result of manual answer will be fed back to machine answering module to improve the capacity of machine answering module. According to the functional requirements of the answering system and the actual retrieval requirements of the self-examinees, the Chinese word partitioning device of Lucene, the full-text search engine, is rewritten. In order to construct the resource index, the system uses the subject ontology database and the word base of Lucene to segment the words simultaneously, and then establishes the index for the segmented words. When retrieving resources, the answering system first uses the concepts in the ontology library to segment the contents entered by the user, and then, according to these segmented keywords, looks up the relevant contents in the indexes of each knowledge base of the system, and then, Then the user's input is segmented by using Lucene's own word partitioning device, and then the key words after segmentation are used to search the relevant contents in the index database. Finally, the retrieval system combines the results of the two searches. The results of the concept lookup using the subject ontology library are ranked first, and the resources found by the keywords partitioned by the Luence particifier are presented to the searcher at the end.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP311.52

【参考文献】