计算机与数字工程2024，Vol.52Issue(10)：2991-2995,3147,6.DOI:10.3969/j.issn.1672-9722.2024.10.024
基于Word2vec与注意力机制的情感分析研究
Research on Sentiment Analysis Based on Word2vec and Attention Mechanisml
摘要
Abstract
In allusion to the problems of inaccurate capture of keyword features and incomplete extraction of local sentiment features by traditional sentiment analysis models,resulting in poor classification results,a sentiment analysis model based on TW-BiLSTM-ATT is proposed.Through the improvement of TF-IDF and the combination with Word2vec,the weight feature is inte-grated into the word vector to improve the ability to capture key information.The position feature of the word vector is integrated into the attention mechanism,the model can focus on the words around the target vocabulary,and then extract the emotional features more comprehensively.The comparative experimental results show that the TW-BiLSTM-ATT model has better classification perfor-mance than similar models in processing sentiment analysis tasks.关键词
Word2vec/TF-IDF/BiLSTM/Attention/情感分析Key words
Word2vec/TF-IDF/BiLSTM/Attention/sentiment analysis分类
信息技术与安全科学引用本文复制引用
任伟建,徐海杰,康朝海,霍凤财,任璐,张永丰..基于Word2vec与注意力机制的情感分析研究[J].计算机与数字工程,2024,52(10):2991-2995,3147,6.基金项目
国家自然科学基金项目(编号:61933007,61873058)资助. （编号:61933007,61873058）
