摘要
微博,作为社交网络的成员之一,以易操作、传播快等特点成为社交网络中的用户发表观点,分享新鲜事的一个重要平台。无论是微博用户所发微博文本信息,还是用户公布在社交媒体中的属性信息,都可能透露出商业中用户的重要属性——消费能力。研究用户的消费能力能够帮助企业针对每个用户的消费能力制定更加有效的商业策略,增大商家投放广告的效率。如今,基于微博的用户画像工作、微博中消费意图识别和商品推荐问题已经成为短文本自然语言处理领域的研究热点。然而,已有工作主要是对用户未公开的基本属性进行预测,或基于微博文本对用户的消费意图进行识别,很少有工作研究微博用户的消费能力。因此,本文主要对微博用户所发微博文本和属性信息与用户的消费能力之间的关系进行研究。首先通过用户链指的方式获取到微博用户的京东账号信息,以京东账号会员等级将消费能力分为高、中、低3个等级。然后将微博文本向量化处理后与用户属性联合在一起,经过特征选择后作为输入训练模型,预测用户的消费能力。
As a member of social networks, microblog becomes an important platform where people like to express their views andshare what happened to them recently due to its simple operation and quick propagation. Either the texts users release or the profilesusers publish on the social media, may reveal the important information in business such as consumption capability. Study on theconsumption capability of users will help companies make more effective business strategies and increase the efficiency of businessadvertising. Today, social media based user profile modeling, consumption intention recognition and product recommendation inmicroblog have become a new research topic in the field of short text natural language processing. However, existing work is mainlyfocused on the prediction of the profiles users didn't publish, or consumption intention recognition based on microblog texts. Theresearch on user's consumption capability is still rare by now. Therefore, this paper focuses on the connection between microblogtexts, user profiles and consumption capability. First, the paper obtains the microblog user 's JingDong account by user link anddivides the consumption level into three levels according to JingDong member level. Then, the paper combines the microblog textswhich are processed with vector quantization and user profiles. After feature selection, the paper regards the information as the inputsof training model and predict user's consumption capability by using the trained model.
作者
裘实
刘挺
QIU Shi;LIU Ting(School of Computer Science and Technology,Harbin Institute of Technology,Harbin 150001,China)
出处
《智能计算机与应用》
2018年第6期180-183,共4页
Intelligent Computer and Applications
关键词
社交媒体
消费能力
微博文本
用户属性
social media
consumption capability
Weibo text
user properties