基于LDA的冷链农产品电商在线评论的情感分析
|
摘要
为促进我国冷链农产品电商企业的发展,提高消费者对商品的满意度和购买欲望,本文从消费者的角度出发,对冷链农产品电商企业的用户评价进行了情感分析,首先依存句法关系制定3种规则提取情感单元,再结合词的词性和依存关系制定情感计算规则,将情感单元分为好评和差评,接着通过LDA主题模型分别对好评与差评进行聚类,根据评价指数S的最大值来确定主题数,最后对聚类的结果进行分析,找出消费者不满意的地方并加以改进。结果表明,提高用户体验、吸引更多消费者,需要从商品的分拣加工、冷链设施的改善、运输配送路程等方面进行优化。本文的研究为整个冷链农产品电商企业的进一步发展提供了借鉴意义。
In order to promote the development of China's cold chain agricultural products e-commerce enterprises and increase consumer satisfaction and purchase desires, this article starts from the perspective of consumers and conducts sentiment analysis on user evaluation of cold chain agricultural products e-commerce enterprises. Relationships formulate three rules to extract emotion units, and then formulate sentiment calculation rules based on the part of speech and dependency relationships of words. The emotional units are divided into positive and negative evaluations. Then the LDA topic model is used to cluster praises and bad reviews, according to the evaluation function. The minimum value of Perplexity is used to determine the number of topics. Finally, the results of clustering are analyzed to find out where the consumers are not satisfied and improve. The experimental results show that in order to improve the user experience and attract more consumers, we must optimize from the aspects of sorting and processing of goods, improvement of cold chain facilities, transportation and distribution distances, and so on, which will further the entire cold chain agricultural product e-commerce business. Development provides lessons for reference.
In order to promote the development of China's cold chain agricultural products e-commerce enterprises and increase consumer satisfaction and purchase desires, this article starts from the perspective of consumers and conducts sentiment analysis on user evaluation of cold chain agricultural products e-commerce enterprises. Relationships formulate three rules to extract emotion units, and then formulate sentiment calculation rules based on the part of speech and dependency relationships of words. The emotional units are divided into positive and negative evaluations. Then the LDA topic model is used to cluster praises and bad reviews, according to the evaluation function. The minimum value of Perplexity is used to determine the number of topics. Finally, the results of clustering are analyzed to find out where the consumers are not satisfied and improve. The experimental results show that in order to improve the user experience and attract more consumers, we must optimize from the aspects of sorting and processing of goods, improvement of cold chain facilities, transportation and distribution distances, and so on, which will further the entire cold chain agricultural product e-commerce business. Development provides lessons for reference.
引文
[1] 胡建淼.我国生鲜农产品冷链物流发展存在的问题与对策[J].改革与战略,2017,33(5):82-84,93.
[2] 蔡晓莹.电子商务环境下生鲜宅配模式改进研究[J].商业经济研究,2018(7):89-90.
[3] 童灿,钱春桃.基于SWOT框架的江苏省农产品电商行业分析[J].安徽农业科学,2015,43(29):360-363.
[4] 崔婧.生鲜电商差异化生存[J].中国经济和信息化,2013(14):52-53.
[5] 邱云飞,陈艺方.基于词性特征与句法分析的商品评价对象提取[J].计算机工程,2016,42(7):173-180.
[6] Khan,Aurangzeb.Sentiment classification using sentence-level semantic orientation of opinion terms from blogs [C].USA:2011 National Postgraduate Conference-Energy and Sustainability,2011.
[7] Chen Tao.A sentence-level sparse gamma topic model for sentiment analysis [J].Lecture Notes in Computer Science,2018,10:316,312.
[8] 王娟,曹树金,谢建国.基于短语句法结构和依存句法分析的情感评价单元抽取[J].情报理论与实践,2017,40(3):107-113.
[9] LIU Kang,XU Liheng.Extracting opinion targets and opinion words from online reviews with graph co-ranking [C].USA:the 52nd Annual Meeting of the ACL,2014:314-324.
[10] Poria S,Cambria E.Aspect extraction for opinion mining with a deep convolutional neural network[J].Knowledge-Based Systems,2016,108:42-49.
[11] 高凯,李思雨.基于微博的情感倾向性分析方法研究[J].中文信息学报,2015,29(04):40-49.
[12] Collobert R.Natural language processing(almost) from scratch[J].The Journal of Machine Learning Research,2011,12:2493-2537.
[13] 王少鹏,彭岩,王洁.基于LDA的文本聚类在网络舆情分析中的应用研究[J].山东大学学报:理学版,2014,49(09):129-134.
[14] 李慧宗.社会化标注环境下的标签聚类方法研究[D].合肥:合肥工业大学,2016:42-44,103.
[15] 薛晶晶,朱占峰.农产品电商物流末端配送问题分析[J].中国商论,2015(26):89-91.
[16] 孙亚娟.基于神经网络的电商生鲜农产品购买频率评价研究[J].商业经济研究,2018(5):131-133.
[17] 高恺,盛宇华.区域性农产品电商平台使用意向影响因素实证研究[J].中国流通经济,2018,32(01):67-74.
[2] 蔡晓莹.电子商务环境下生鲜宅配模式改进研究[J].商业经济研究,2018(7):89-90.
[3] 童灿,钱春桃.基于SWOT框架的江苏省农产品电商行业分析[J].安徽农业科学,2015,43(29):360-363.
[4] 崔婧.生鲜电商差异化生存[J].中国经济和信息化,2013(14):52-53.
[5] 邱云飞,陈艺方.基于词性特征与句法分析的商品评价对象提取[J].计算机工程,2016,42(7):173-180.
[6] Khan,Aurangzeb.Sentiment classification using sentence-level semantic orientation of opinion terms from blogs [C].USA:2011 National Postgraduate Conference-Energy and Sustainability,2011.
[7] Chen Tao.A sentence-level sparse gamma topic model for sentiment analysis [J].Lecture Notes in Computer Science,2018,10:316,312.
[8] 王娟,曹树金,谢建国.基于短语句法结构和依存句法分析的情感评价单元抽取[J].情报理论与实践,2017,40(3):107-113.
[9] LIU Kang,XU Liheng.Extracting opinion targets and opinion words from online reviews with graph co-ranking [C].USA:the 52nd Annual Meeting of the ACL,2014:314-324.
[10] Poria S,Cambria E.Aspect extraction for opinion mining with a deep convolutional neural network[J].Knowledge-Based Systems,2016,108:42-49.
[11] 高凯,李思雨.基于微博的情感倾向性分析方法研究[J].中文信息学报,2015,29(04):40-49.
[12] Collobert R.Natural language processing(almost) from scratch[J].The Journal of Machine Learning Research,2011,12:2493-2537.
[13] 王少鹏,彭岩,王洁.基于LDA的文本聚类在网络舆情分析中的应用研究[J].山东大学学报:理学版,2014,49(09):129-134.
[14] 李慧宗.社会化标注环境下的标签聚类方法研究[D].合肥:合肥工业大学,2016:42-44,103.
[15] 薛晶晶,朱占峰.农产品电商物流末端配送问题分析[J].中国商论,2015(26):89-91.
[16] 孙亚娟.基于神经网络的电商生鲜农产品购买频率评价研究[J].商业经济研究,2018(5):131-133.
[17] 高恺,盛宇华.区域性农产品电商平台使用意向影响因素实证研究[J].中国流通经济,2018,32(01):67-74.