摘要
随着电商的飞速发展,越来越多的人选择通过网络来购物,随之而来的是消费者在各大购物网站上发表的大量评价信息。为了从中快速识别出关键的评价信息,本文基于协同训练算法(Co-training),利用条件随机场(CRF)模型,设计了一种自动识别关键的评价信息的方法。与2014COAE评测对比,该方法的效果优于评测平均水平,验证了该方法的有效性。
With the rapid development of electronic commerce,more and more people choose online shopping.Consumers have published a lot of comments on major shopping sites.This paper designs a automatic method to identify key review information based on co-training by CRF model.Compared with 2014COAE,the result of this method is better than average level,and its effectiveness is confirmed.
作者
乔磊
Qiao Lei(Shanxi Dazhong Electronic Information Industry Group Co., Ltd.,Taiyuan Shanxi 030024,China)
出处
《山西电子技术》
2020年第4期65-67,70,共4页
Shanxi Electronic Technology
