基于改进TFIDF算法的情感分析模型研究  

Research on Emotion Analysis Model Based on Improved TFIDF Algorithm

作　　者：季旺 夏振宇 JI Wang;XIA Zhenyu(School of Computer,Jiangsu University of Science and Technology,Zhenjiang 212003)

机构地区：[1]江苏科技大学计算机学院,镇江212003

出　　处：《计算机与数字工程》2022年第8期1671-1675,共5页Computer & Digital Engineering

摘　　要：随着电商行业的蓬勃发展,网上购物逐渐取代线下商店成为最受欢迎的购物方式之一。因此从海量的商品评价中挖掘出有用的信息,对顾客购买商品和商家提高服务质量具有重要的意义。在深度学习背景下,论文在关键词提取中对TF-IDF算法进行改进,主要是关键词权重的优化。实验结果表明,基于论文改进的算法构建的商品评论模型比传统的模型分类效果好。With the vigorous development of ecommerce market,shopping on the internet has substituted for offline shops by degrees,becoming one of the most fashionable ways to shop.Therefore,mining useful information from a large number of commodity evaluation is of great significance for consumers to buy goods and businesses to improve service quality.In the context of deep learning,the paper improves the TF-IDF algorithm in keyword extraction,mainly the optimization of keyword weights.Test results indicate that in view of the improved algorithm of this article,the merchandise evaluate pattern is better than the traditional model classification.

关 键 词：商品评价 深度学习 关键词提取 TF-IDF 权重 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]