-
摘要在电商蓬勃发展的背景下，以用户为中心的商品评论信息不断增加。对海量非结构化的商品评论信息进行观点抽取，能够帮助企业和消费者快速获得有效信息。本文将CRFs模型和本体的方法相结合，对商品评论信息中的评价对象和评价词抽取进行了研究，提出基于评价对象、评价词自身特征的观点抽取模型。同时归纳整理商品评论中观点的生成模式，以达到对商品评论信息中评价观点抽取的目的。在商品评论信息观点抽取的基础上，依照评价对象和评价词在本体中的不同特点对观点进行分类。本文以汽车商品为例，依照提出的观点抽取和观点分类模型进行了实验，发现该模型对常用的、结构简单的评论句抽取的效率较高，对于句式结构复杂的评论句抽取的效率较差，并由此进一步分析模型产生误差的原因。27016
- 上一篇：关键词共现网络中的主题突变识别与分析研究
- 下一篇：网络舆情链式效应结构关系及其演化规律分析
关键词 观点抽取 商品评论信息 CRFs模型 本体
毕业论文外文摘要
Title Research on the Extraction and Classification of Commodity Reviews Information
Abstract
Under the background of the vigorous development of e-commerce, the commodity review information centered on customer continues to increase. It could help the enterprises and customers get the effective information quickly that extracting massive unstructured commodity reviews information. This paper combined CRFs model and ontology, proposed an opinion extraction model based on the features of evaluation object and evaluation word, to extract evaluation object and evaluation word. Meanwhile, we concluded the ways of opinion expression, to help evaluation opinion extraction from user reviews. Based on the opinion extraction, opinions were classified according to their evaluation objects and evaluation words.
This paper chose the car product as an example. According to the model, we made the experiment about opinion extraction and opinion classification. In this paper, it took automobiles for examples and conducted the opinion extraction in terms of the model above. It is found that the model shown higher extraction efficiency in the reviews with simple structure than the reviews with complex structure. Then the underlying reasons about the efficiency differences were examined.
Keywords Opinion Extraction Product Reviews Information CRFs Model Ontology
目 次
1. 绪论 1
1.1 研究背景 1
1.2 研究意义 1
1.3 观点抽取综述 1
1.4 观点分类综述 3
1.5 论文研究内容 3
2. 方法模型研究 5
2.1 相关方法与理论 5
2.1.1 条件随机场模型 5
2.1.2 本体 6
2.1.3 句法分析 6
2.2 任务需求分析 7
2.3 方法模型及规则设计 7
2.3.1 观点抽取模型 7
2.3.2 观点生成规则 9
2.3.3 观点分类模型 10
2.4 商品评论本体 10
2.4.1 评价对象类设计 11
2.4.2 评价词类设计 12
2.4.3 意愿词类设计 12
2.5 模型应用领域 14
3. 系统实现 15
3.1 系统需求 15
3.2 实现工具 15
3.3 系统模块 15
3.3.1 评论信息抓取模块 15
3.3.2 观点抽取及分类模块 18
3.3 系统分析 19
4. 实验及结果分析 20