基于机器阅读与文本标注的语文阅读理解试题智能化管理系统

Intelligent Management System of Chinese Reading Comprehension Test Based on Machine Reading and Text Annotation

本文提出了一种解决阅读理解测试中的答题问题的方法。系统以语文文档作为输入,回答有关该文档的多项选择问题,通过使用Lucene信息检索引擎通过附加的自动语言处理(例如词干、回指解析和词性标记)进行信息提取。通过比较Lucene为每个问题相对于其候选答案检索的文本来验证答案,因此,执行基于文本蕴含的验证。通过实验评估验证使用在语文阅读理解中广泛使用的两种语料库提出的方法的质量。结果表明,所提出的系统以33-37%的百分比选择了给定问题的正确答案。

This article proposes a method to solve the answering questions in the reading comprehension test.The system takes a Chinese document as input,answers multiple choice questions about the document,and uses the Lucene information retrieval engine to extract information through additional automatic language processing(such as stemming,anaphora analysis,and part-of-speech tags).The answer is verified by comparing the text retrieved by Lucene for each question relative to its candidate answer,and therefore,verification based on the implication of the text is performed.The quality of the methods proposed by the two corpora that are widely used in Chinese reading comprehension is verified through experimental evaluation.The results show that the proposed system selects the correct answer for a given question in a percentage of 33-37%.