摘要
互联网的飞跃发展,既孕育着机遇,同时也带来了前所未有的挑战。网络舆情的特点使其成为一把双刃剑。为此,本文通过对微博文本的获取与处理,得到关于该微博热门话题的基于时间序列的离散数据序列,然后采用万有引力算法优化的RBF神经网络对微博舆情进行预测。通过微博舆情的时间序列进行实证研究,在预测性能上与现有的预测模型进行对比,证明该模型在该预测领域的可行性和有效性。
The rapid development of internet,while having brought new good fortune to us,has also broughtunprecedented challenges.The characteristics of network public opinion make it a double-edged sword.There?fore,this article gets discrete data sequence of micro-blog hot topics based on time series by gaining and pro?cessing micro-blog text,and then makes a prediction for micro-blog public opinion using RBF neural networkoptimized by gravitational algorithm.Empirical study is carried out through the time series of micro-blog publicopinion,and the optimized model is compared with the existing prediction model from the prediction perfor?mance,indicting this model has feasibility and effectiveness in this prediction field.
作者
王阳
Wang Yang(Lanzhou Jiaotong University, Lanzhou Gansu 730070)
出处
《创新科技》
2016年第12期32-35,共4页
Innovation science and technology
基金
河南省高等学校重点科研项目"社会舆情监控与有效引导对策研究"(15A520042)
河南省教育厅人文社会科学研究重点项目"科学利用网络舆情促进服务型政府建设研究"(2015-ZD-137)
关键词
RBF神经网络
微博舆情
万有引力算法
预测模型
RBF neural network
micro-blog public opinion
gravitational algorithm
prediction model
