摘要
针对网络舆情发展过程的复杂性和多成分的特点,本文提出一种基于小波分析和人工神经网络的网络舆情建模和预测方法。利用小波变换将舆情发展过程分解到不同的尺度层面上,形成舆情发展过程的整体趋势、细节等部分,然后通过人工神经网络对每个部分进行建模并最终达到预测的目的,并与其他预测模型进行对比实验。结果表明,相对于其他模型,本文提出的小波-人工神经网络预测模型具有较高的精确度和稳定性。
According to the characteristics of complexity of network public opinion development process and multicomponents, this paper presents a method of network modeling and forecasting based on wavelet analysis and artificialneural network. The public opinion development process, which contains the overall trend and the details of thedevelopment process, is decomposited to different levels of the scale by wavelet function, and then through the artificialneural network for each level is modeled and ultimately achieve the purpose of forecasting, and is compared with otherforecasting models. The results show that, compared with other models,the prediction model based on the wavelet and theartifical neural network has higher accuracy and stability.
出处
《情报科学》
CSSCI
北大核心
2016年第4期40-42,47,共4页
Information Science
关键词
网络舆情
小波分析
人工神经网络
internet public opinions
wavelet analysis
artifical neural networks