成都大学学报（自然科学版）2025，Vol.44Issue(3)：270-279,10.DOI:10.3969/j.issn.1004-5422.2025.03.007
基于多代理的情感分类及预测模型研究
Research on Sentiment Classification and Prediction Model Based on Multiple Agents
摘要
Abstract
This article addresses the issue of low accuracy in the sentiment classification and prediction of community evaluations in social network big data.On the basis of the traditional LSTM sentiment classifi-cation model based on deep learning,the original model is optimized on public datasets through BIGRU,CNN,loss functions and cross-attention mechanisms,and the sentiment classification model BICGRU is proposed.The results of the sentiment classification model processing the dataset are visualized and ana-lyzed through the social network sentiment prediction model netlogo sentiment prediction model construc-ted by netlogo,and complex network and neural network technologies are used to study information dis-semination,opinion formation and decision-making processes.The experimental results on public datasets show that the improved model can more accurately capture the feature differences between sentiment po-larities;the social network sentiment prediction model can successfully simulate information dissemina-tion,opinion formation and decision-making processes,and has high scalability and adaptability,which can be used to analyze various community sentiment analysis.关键词
深度学习/BICGRU/netlogo/社区情感预测模型Key words
deep learning/BICGRU/netlogo/netlogo sentiment prediction model分类
信息技术与安全科学引用本文复制引用
周艳玲,沈晓轩..基于多代理的情感分类及预测模型研究[J].成都大学学报（自然科学版）,2025,44(3):270-279,10.基金项目
安徽省大学生创新创业项目(1602652994510262272、1602651122336534528、1602648829042102272) （1602652994510262272、1602651122336534528、1602648829042102272）
2022年国家级大学生创新创业项目(202211059091、202211059021) （202211059091、202211059021）
安徽省质量工程项目(2021xxkc158) （2021xxkc158）
安徽省教学研究项目(2024jyxm0330) （2024jyxm0330）
