基于小波变换与差分自回归移动平均模型的微博话题热度预测
发布时间：2018-11-20 05:17
【摘要】：研究话题热度预测问题对于网络广告传播效应最大化、网络舆论引导与控制等具有重要意义.首先,根据用户关系及话题因素计算用户影响力,进而定义话题影响力.然后,基于老化理论并考虑话题影响力和话题相关微博数定义话题能量值,量化话题热度.最后,提出基于小波变换与差分自回归移动平均模型的微博话题热度预测方法,以此预测话题热度(能量值)及话题能量峰值.实验表明,文中方法可有效预测话题热度及峰值,具有较低的残差和遗漏率.
[Abstract]:The research of topic heat prediction is of great significance to maximize the effect of network advertising communication and guide and control network public opinion. First, the user influence is calculated according to the user relationship and topic factors, and then the topic influence is defined. Then, based on aging theory and considering topic influence and topic correlation Weibo number, we define topic energy value and quantify topic heat. Finally, based on wavelet transform and differential autoregressive moving average model, Weibo topic heat prediction method is proposed to predict topic heat (energy value) and topic energy peak value. The experimental results show that the proposed method can effectively predict the heat and peak of the topic, and has low residual and omission rates.
【作者单位】： 福州大学数学与计算机科学学院;福州大学福建省网络计算与智能信息处理重点实验室;
【基金】：国家自然科学基金项目(No.61300104,61370210,61103175) 福建省自然科学基金项目(No.2013J01232) 福建省教育厅重点项目(No.JK2012003) 福建省科技创新平台项目(No.2009J1007) 福建省科技厅产学重大项目(No.2014H6014)资助
【分类号】：TP393.092
[Abstract]:The research of topic heat prediction is of great significance to maximize the effect of network advertising communication and guide and control network public opinion. First, the user influence is calculated according to the user relationship and topic factors, and then the topic influence is defined. Then, based on aging theory and considering topic influence and topic correlation Weibo number, we define topic energy value and quantify topic heat. Finally, based on wavelet transform and differential autoregressive moving average model, Weibo topic heat prediction method is proposed to predict topic heat (energy value) and topic energy peak value. The experimental results show that the proposed method can effectively predict the heat and peak of the topic, and has low residual and omission rates.
【作者单位】： 福州大学数学与计算机科学学院;福州大学福建省网络计算与智能信息处理重点实验室;
【基金】：国家自然科学基金项目(No.61300104,61370210,61103175) 福建省自然科学基金项目(No.2013J01232) 福建省教育厅重点项目(No.JK2012003) 福建省科技创新平台项目(No.2009J1007) 福建省科技厅产学重大项目(No.2014H6014)资助
【分类号】：TP393.092
【参考文献】
相关期刊论文 前4条
1 邹柏贤,刘强;基于ARMA模型的网络流量预测[J];计算机研究与发展;2002年12期
2 洪飞;吴志美;;基于小波的多尺度网络流量预测模型[J];计算机学报;2006年01期
3 路荣;张e，
本文编号：2343872
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2343872.html
