摘要
在不同的电商平台中,部分用户购买商品后会发表评论信息,以此来反馈购买商品的态度。对用户的商品评论数据进行挖掘与分析,有利于商户与生产企业预估商品销量、改进商品品质而言具有潜在的应用价值。因此本文主要工作是采用改进的Word2Vec词嵌入模型,将整个情感词典进行扩展分析,以改进情感词典的电商平台适用性,确定用户真实的情感倾向,再将其与情感词极性分类算法有效结合,最终提升整体算法的情感分类性能,实验结果表明,所提Conv1d-Word2Vec模型相较于传统模型具有更优的情感识别效果。
In different e-commerce platforms, some users will post comments after purchasing products to feedback their attitude towards the products. Mining and analyzing product review data of users has good practical value for enterprises. In this context, the main work of this paper is to use the improved Word2Vec word embedding model, to expand and analyze the entire sentiment dictionary, building a good sentiment dictionary of e-commerce business platform, and determining users’ real emotional tendencies, it is combined with word polarity algorithm effectively. Finally, the emotion analysis and improvement of the whole algorithm are carried out. The experimental results that Conv1d-Word2Vec model has better effect on emotion recognition than traditional model.
作者
李昕昊
LI Xinhao(University of Shanghai for Science and Technology,Shanghai 200093)
出处
《软件》
2022年第12期100-104,共5页
Software
关键词
词嵌入
情感分类
一维卷积
词极性算法
Word2Vec
sentiment classification
conv1d
word polarity algorithm
作者简介
李昕昊(1995-),男,陕西西安人,硕士研究生,研究方向:系统工程。
