互联网的快速发展给各大电商平台和生产厂家带来机遇的同时也带来了挑战。用户在互联网上购物的同时,产生了海量的评论数据,而在这些评论文本中包含着许多有价值的潜在信息,因此通过对商品评论信息的分析,不仅能让企业掌握更多自身产品...互联网的快速发展给各大电商平台和生产厂家带来机遇的同时也带来了挑战。用户在互联网上购物的同时,产生了海量的评论数据,而在这些评论文本中包含着许多有价值的潜在信息,因此通过对商品评论信息的分析,不仅能让企业掌握更多自身产品和服务中的具体细节信息,同时能够进一步分析用户的消费行为,从本质上发现用户的需求偏好,推进企业实施科学经营决策。本文的研究对象是笔记本电脑,使用爬虫技术获取联想拯救者Y9000P的用户评论,对数据进行预处理、分词与词性标注,采用余弦相似度的方法进行主题数寻优,确定主题数后建立隐藏式狄利克雷模型(Latent Dirichlet Allocation),挖掘用户高频关注的产品属性,用词典匹配的方法匹配情感词,进行情感倾向分析,得到用户对产品的意见、态度、购买偏好、购买习惯以及购买动机。The rapid development of the Internet has brought opportunities as well as challenges to major e-commerce platforms and manufacturers. While users are shopping on the Internet, they generate a large amount of comment data, and these comment texts contain many valuable potential information. Therefore, through the analysis of commodity comment information, enterprises can not only grasp more specific details of their own products and services, but also further analyze users’ consumption behavior, discover users’ demand preferences in essence, and promote enterprises to implement scientific management decisions. The research object of this paper is notebook computer. It uses crawler technology to obtain user comments of Lenovo savior Y9000P, preprocesses the data, segment and label the part of speech, uses cosine similarity method to optimize the number of topics, establishes LDA model after determining the number of topics, excavates the product attributes that users pay high attention to, uses dictionary matching method to match emotional words, carries out emotional tendency analysis, and obtains users’ opinions, attitudes, purchase preferences, habits and motivations.展开更多
文摘互联网的快速发展给各大电商平台和生产厂家带来机遇的同时也带来了挑战。用户在互联网上购物的同时,产生了海量的评论数据,而在这些评论文本中包含着许多有价值的潜在信息,因此通过对商品评论信息的分析,不仅能让企业掌握更多自身产品和服务中的具体细节信息,同时能够进一步分析用户的消费行为,从本质上发现用户的需求偏好,推进企业实施科学经营决策。本文的研究对象是笔记本电脑,使用爬虫技术获取联想拯救者Y9000P的用户评论,对数据进行预处理、分词与词性标注,采用余弦相似度的方法进行主题数寻优,确定主题数后建立隐藏式狄利克雷模型(Latent Dirichlet Allocation),挖掘用户高频关注的产品属性,用词典匹配的方法匹配情感词,进行情感倾向分析,得到用户对产品的意见、态度、购买偏好、购买习惯以及购买动机。The rapid development of the Internet has brought opportunities as well as challenges to major e-commerce platforms and manufacturers. While users are shopping on the Internet, they generate a large amount of comment data, and these comment texts contain many valuable potential information. Therefore, through the analysis of commodity comment information, enterprises can not only grasp more specific details of their own products and services, but also further analyze users’ consumption behavior, discover users’ demand preferences in essence, and promote enterprises to implement scientific management decisions. The research object of this paper is notebook computer. It uses crawler technology to obtain user comments of Lenovo savior Y9000P, preprocesses the data, segment and label the part of speech, uses cosine similarity method to optimize the number of topics, establishes LDA model after determining the number of topics, excavates the product attributes that users pay high attention to, uses dictionary matching method to match emotional words, carries out emotional tendency analysis, and obtains users’ opinions, attitudes, purchase preferences, habits and motivations.
