摘要
随着社交网络的日益庞大,各类评论信息产生的渠道和数量也飞速增长,通过人工阅读所有评论来了解口碑情况变得日益困难,所以构建一个精准的口碑评论分值预测模型对商家和用户来说都显得日益重要。文章旨在对真实口碑评论数据进行分析挖掘和多维度特征提取,并构建一个基于多特征的加权融合模型对口碑评论的评分值进行预测。通过实验证明,在当前数据基础上,该模型可以有效地对口碑评论进行预测,相比传统方法,效果更好。
With the growing number of social networks, the channels and numbers of various types of commentary information are alsogrowing rapidly. It is becoming increasingly difficult to understand the word-of-mouth situation by manually reading all the comments,so construct a precise word-of-mouth commentary score prediction model for merchants and users. It seems to be increasingly important.This paper aims to analyze and mine the real word-of-mouth commentary data and multi-dimensional feature extraction, and construct aweighted fusion model based on multi-features to predict the scores of word-of-mouth comments. Experiments show that the model caneffectively predict word-of-mouth comments based on current data, and the effect is better than traditional methods.
作者
郭瑞祥
左彬靖
杜成喜
肖明
王杰
Guo Ruixiang;Zuo Binjing;Du Chengxi;Xiao Ming;Wang Jie(School of Automation, Guangdong University of Technology, Guangzhou 510006, China;Guangdong Modern AudiovisualInformation Engineering Technology Research Center, Guangzhou University, Guangzhou 510006, China)
出处
《无线互联科技》
2019年第1期68-70,共3页
Wireless Internet Technology
基金
广东省科技计划项目
项目编号:2015B010131014
2017B010125002
关键词
口碑评论
特征提取
机器学习
word-of-mouth commentary
feature extraction
machine learning
