当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术21年15期

基于 SVM 的商品评论情感研究
韩美玉
(宁夏大学新华学院,宁夏 银川 750021)

摘  要:为充分挖掘用户对网购商品的评论信息,为消费者的购买决策提供参考,同时帮助商家改进自身产品、提高市场竞争力,文章提出一种基于支持向量机(SVM)的细粒度商品评论情感分析方法。首先,使用 Python 中的网络爬虫获取京东某品牌的冰箱评论信息作为语料库并对其进行预处理,将语料数据分为训练集和测试集;接着,通过特征选择对词汇集做降维处理并使用支持向量机(SVM)的算法对商品评论信息进行情感分类;最后,统计包含每个基本属性和其扩充的特征词集的正面评论个数及负面评论个数,分析并给出结论。


关键词:SVM;文本情感研究;商品评论情感研究



DOI:10.19850/j.cnki.2096-4706.2021.15.032


基金项目:宁夏大学新华学院科学研究基 金项目资助项目(20XHKY10)


中图分类号:TP181                                          文献标识码:A                                     文章编号:2096-4706(2021)15-0122-03


Research on Sentiment of Product Review Based on SVM

HAN Meiyu

(Xinhua College of Ningxia University, Yinchuan 750021, China)

Abstract: In order to fully mine the online shopping comment information of a certain product, provide reference for consumers’ purchase decision, and help merchants improve their own products and enhance market competitiveness, this paper proposes a fine-grained sentiment analysis method of product comment based on support vector machine (SVM). Firstly, the network crawler in Python is used to obtain the comments of a refrigerator of Jingdong brand as a corpus. The corpus data is divided into training set and test set, and the corpus is preprocessed. Secondly, feature selection is used to reduce the dimension of the word collection and support vector machine (SVM) algorithm is used to classify the product review information. Finally, the number of positive comments and negative comments of each basic attribute and it’s expanded feature word set are counted, analyzed and concluded.

Keywords: SVM; text emotion research; goods comments emotion research


参考文献:

[1] PANG B,LEE L,VAITHYANATHAN S.Thumbs up Sentiment classification using machine learning techniques [C]// Computation and Language,may28,2002.

[2] 李明,胡吉霞,侯琳娜,等 . 商品评论情感倾向性分析 [J]. 计算机应用,2019,39(S2):15-19.

[3] 宗成庆,夏睿,张家俊 . 文本数据挖掘 [M]. 北京:清华大学出版社,2019:65-67.

[4] 肖江,王晓进 . 基于 SVM 的在线商品评论的情感倾向性分析 [J]. 信息技术,2016(7):172-175.

[5] 刘若雨 . 基于电商评论文本的用户情感分析 [J]. 现代信息科技,2021,5(4):85-87+92.


作者简介:韩美玉(1991—),女,汉族,宁夏银川人,助教,硕士研究生,研究方向:文本情感研究。