摘要
建立有效的用户行为预测模型,准确地预测用户的上网行为,是当前网络主动管理地关键,传统的Markov模型是一种简单而有效的预测模型,但它存在测准确率低、预测覆盖率低以及存储复杂度高等缺点。提出了基于加权马尔可夫链模型,通过分析用户行为特征和最优状态分类的方法,预测网络用户行为。最后通过实验结果表明了该模型的可行性和实用性。
Modeling users＇ behavior and accurate predictions of users online behavior is the key to the current network management.The traditional Markov model is simple and practical, but it gives low prediction accuracy and coverage rate, as well as requires high space complexity. The model based on weighted Markov chain is presented, it can predict the behavior of network user by analyzing the user behavior state and the classification of optimal state. In addition, its feasibility and practicality are validated.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第10期3334-3337,3418,共5页
Computer Engineering and Design
基金
国家教育部科学技术研究重点基金项目(208146)
关键词
马尔可夫链模型
用户行为
行为分析
网络流量
预测模型
Markov chain model
user behavior
behaviors analysis
network traffic
prediction model
