摘要
在线商品评论是促进电子商务健康发展的重要内容,然而虚假评论的存在严重扰乱了公平、公正的市场竞争秩序。因此,评论的真伪鉴别是亟需解决的重要问题。本文在采集含标签评论集的基础上,通过文本嵌入进行特征提取,训练了一个长短期记忆神经网络模型。测试结果显示,此分类器能够有效地识别虚假评论。
Online commodity review is an important content to promote the healthy development of e-commerce.However,the existence of false comments has seriously disturbed the fair and just market competition order,so the identification of the authenticity of comments is an important problem that needs to be solved urgently.In this paper,a neural network model of short-term and long-term memory is trained based on the collection of tagged comment sets and feature extraction through text embedding.The test results show that this classifier can effectively identify false comments.
作者
林婧雯
李建敦
王赢胜
丁嘉华
罗啸驰
LIN Jingwen;LI Jiandun;WANG Yingsheng;DING Jiahua;LUO Xiaochi(School of Electronic Information Engineering,Shanghai Dianji University,Shanghai,China,201306)
出处
《福建电脑》
2022年第8期10-13,共4页
Journal of Fujian Computer
基金
上海市大学生创新创业训练计划(No.B1-0224-21-004-00-171)资助。
关键词
虚假评论
自然语言
特征提取
深度模型
长短期记忆模型
Untruthful Reviews
Natural Language
Feature Extraction
Deep Models
Long Short-Term Memory Model