摘 要

随着时代的不断发展与科技的不断进步，人工智能已经进入到了我们的生活当中，成为我们生活中不可或缺的组成部分，汉字作为我们的母语也迎来了崭新的发展局面。语义识别系统的研究说到底其实就是汉语研究与计算机等高科技结合来完成新技术变革的一项研究。就像是现在的教育发展趋势之一就是教学现代化，比如，网上教学、无纸化考试等，这些技术相对来说虽然还有着一定的发展空间但其实已经趋于成熟。当然，这些网上考试的安排也衍生了新技术，客观题自动阅卷应运而生的同时，主观题自动阅卷系统的研究也一步步提上了日程。为了日后的教育现代化大力发展，主观题自动阅卷系统研究在未来的一段时间里将会引起一定的热潮。

对于主观题比如论述，简答，作文等这些试题答案涉及到了汉字本身的语言特点，用不同的字词句表达出来的内容可能说明了同一个道理，不同字、词、句的提取可能影响到整篇文档的了解，这就涉及到了自然语言处理、语义识别、人工智能等多项技术，同时，要想要做到主观题自动阅卷，还需要通过学生文本与标准文本之间的相似度计算得出学生得分，完成系统研究。

关键词：自然语言处理；自动阅卷；语义识别；语义相似度计算

Abstract

With the continuous development of the times and the continuous progress of science and technology, artificial intelligence has entered into our lives, become an indispensable part of our lives, Chinese characters as our mother tongue also ushered in a new development situation. Semantic recognition system research in the end is actually Chinese research and computer and other high-tech combination to complete a new technology change in a study. Just as one of the current educational trends is the modernization of teaching, such as online teaching, paperless examinations, etc., these technologies are still relatively mature, although there is still some room for development. Of course, the arrangements for these online examinations also derived new technology, objective questions automatically read the volume came into being at the same time, the subjective automatic marking system research step by step on the agenda. In order to develop the modernization of education in the future, the subjective automatic marking system research will cause some upsurge in the future for some time.

For the subjective questions such as discussion, short answer, composition, etc. The answers to these questions related to the Chinese character of the language itself, with different words to express the content may explain the same reason, different words, words, sentence extraction may affect The whole document understanding, which involves the natural language processing, semantic identification, artificial intelligence and many other technologies, at the same time, to be subject to subjective questions automatically read, but also through the student text and standard text similarity between the calculation Get the student score, complete the system research.

Key Words：Natural language processing; automatic marking; semantic recognition; semantic similarity calculation

目 录

第1章 绪论 1

1.1研究背景及意义 1

1.2国内外发展现状 2

1.2.1国内发展现状 2

1.3本文的主要工作 3

1.4章节安排 3

1.5本章小结 4

第2章语义识别系统的介绍 5

2.1信息检索 5

2.2 信息提取 6

2.3 TF-IDF 6

2.4 奇异值分解 7

2.5 本章小结 7

第3章 语义识别核心技术的分析 9

3.1潜在语义分析方法 9

3.2 向量空间模型 9

3.3余弦相似度模型 10

3.4本章小结 11

第4章 分析一个基本的语义识别过程 12

4.1 潜在语义空间的构建 12

4.2 文本预处理 12

4.3 信息检索与关键字提取 12

4.3.1 信息检索 12

4.3.2 关键词提取 12

4.4 特征项选择 13

4.5 文本相似度计算 13

4.6 本章小结 14

第5章总结与展望 15

5.1总结 15

5.2展望 15

参考文献 17

致 谢 18