互联网社交平台舆论趋势预测算法研究  

Research on public opinion trend prediction algorithm on Internet social platform

作　　者：王海兮[1] 吴喆熹 马军[1] WANG Haixi;WU Zhexi;MA Jun(The 30th Institute of China Electronics Technology Corporation,Chengdu 610041,China)

机构地区：[1]中国电子科技集团公司第三十研究所,四川成都610041

出　　处：《应用科技》2023年第6期101-105,共5页Applied Science and Technology

摘　　要：为解决互联网社交平台话题趋势预测不准确的问题,提出基于量化计算的互联网话题趋势预测算法。本文通过分析互联网特定话题下的主题立场文本、评论账户画像、评论贴文内容等多维度特征,设计了一种面向互联网特定话题的舆论趋势预测方法,通过特定话题数据获取、用户观点立场分析、趋势走向判断样本集构建模型,构建的互联网话题趋势预测神经网络模型比传统的舆论趋势预测模型更准确、更全面。本文提出的基于量化计算的互联网话题趋势预测算法,可作为互联网舆情趋势分析研判的重要手段。In order to solve the problem of inaccurate trend prediction on public opinions on the Internet social platform,this paper proposes a trend prediction algorithm on Internet topics based on quantitative calculation.By analyzing multi�dimensional features under specific Internet topic such as text of topic standpoints,portrayal of comment account and contents of comment words,etc.,this paper designs a trend prediction algorithm on public opinions orienting to a specific Internet topic.Through the acquisition of data on specific topics,the analysis data of user views and positions,construction of sample set for trend judgment,the etablished neural network model on trend prediction of Internet topic can be more accurate and comprehensive than the traditional public opinion trend prediction model.The Internet topic trend prediction algorithm based on quantitative calculation can be used as an important means for public opinion trend prediction on the Internet.

关 键 词：舆论分析 舆论趋势 神经网络模型 账号属性特征 内容语义特征 话题立场语义特征 舆论趋势预测 立场变化 

分 类 号：TP302.1[自动化与计算机技术—计算机系统结构]