基于收视行为的互联网电视节目流行度预测模型
发布时间：2018-08-21 14:20
【摘要】：准确预测节目流行度是互联网电视节目系统设计与优化所要解决的关键问题之一。针对现有预测方法存在模型训练时间长、样本数量多、且对突发热点节目流行度预测效果差等问题,该文测量了某互联网电视平台280万用户的60亿条收视行为数据,采用行为动力学分类方法将节目流行度演化过程分为内源临界、内源亚临界、外源临界和外源亚临界4种类型,运用双种群粒子优化的最小二乘支持向量机对每种类型分别构建了一种互联网电视节目流行度预测模型BD3P,并将BD3P模型应用于实际数据测验。实验结果表明,与现有其他方法相比,BD3P模型预测精度可提升17%以上,并能有效缩短预测周期。
[Abstract]:Accurate prediction of program popularity is one of the key problems to be solved in the design and optimization of Internet TV program system. Aiming at the problems of long training time of model, large number of samples and poor prediction effect on the popularity of burst hot spots, this paper measures 6 billion data of viewing behavior of 2.8 million users of an Internet TV platform. The evolution process of program popularity was classified into four types: endogenous critical, endogenous subcritical, exogenous critical and exogenous subcritical. Based on the two-population particle optimization least squares support vector machine (LS-SVM), an Internet TV program popularity prediction model BD3P is constructed for each type, and the BD3P model is applied to the actual data test. The experimental results show that compared with other existing methods, the prediction accuracy of the model can be improved by more than 17%, and the prediction period can be shortened effectively.
【作者单位】： 东南大学计算机科学与工程学院;教育部计算机网络和信息集成重点实验室(东南大学);
【基金】：国家863计划项目(2015AA015603) 江苏省未来网络创新研究院未来网络前瞻性研究项目(BY2013095-5-03) 江苏省“六大人才高峰”高层次人才项目(2011-DZ024)~~
【分类号】：TN949.292
本文编号：2196014
[Abstract]:Accurate prediction of program popularity is one of the key problems to be solved in the design and optimization of Internet TV program system. Aiming at the problems of long training time of model, large number of samples and poor prediction effect on the popularity of burst hot spots, this paper measures 6 billion data of viewing behavior of 2.8 million users of an Internet TV platform. The evolution process of program popularity was classified into four types: endogenous critical, endogenous subcritical, exogenous critical and exogenous subcritical. Based on the two-population particle optimization least squares support vector machine (LS-SVM), an Internet TV program popularity prediction model BD3P is constructed for each type, and the BD3P model is applied to the actual data test. The experimental results show that compared with other existing methods, the prediction accuracy of the model can be improved by more than 17%, and the prediction period can be shortened effectively.
【作者单位】： 东南大学计算机科学与工程学院;教育部计算机网络和信息集成重点实验室(东南大学);
【基金】：国家863计划项目(2015AA015603) 江苏省未来网络创新研究院未来网络前瞻性研究项目(BY2013095-5-03) 江苏省“六大人才高峰”高层次人才项目(2011-DZ024)~~
【分类号】：TN949.292
【相似文献】
相关期刊论文 前1条
1 吴绮;;基于节点流行度的Gnutella路由查询策略[J];科技信息;2009年02期
相关硕士学位论文 前2条
1 张志伟;基站环境下媒体流行度预测与缓存策略研究[D];中国科学技术大学;2016年
2 牛振兴;媒体分发应用中基于内容流行度缓存技术的研究[D];华中科技大学;2006年
，本文编号：2196014
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2196014.html
