摘要
有效地预测互联网用户在浏览过程中的行为方向,尽可能有针对性的提供其感兴趣的信息,有助于提高网站的服务质量,最大限度地留住客户。马尔科夫模型是解决这一问题的一种有效建模方式。但是该模型,尤其是高阶马尔科夫模型的指数级复杂度大大影响了其实际的应用。建立新的模型需要权衡自身的预测准确度与复杂度,尽可能保持前者,同时降低后者。结合马尔科夫模型和有向图特点,提出了一种新的建模方式。新模型具有与马尔科夫模型相同的预测准确度,但是各种时间、空间复杂度相对较低,综合性能良好。
Predicting user＇s directions efficiently and correctly will be of great benefit to the web site and Markov model is a good solution for this subject. But the Markov model especially high order Markov models suffer a lot from their extremely high complexity. New model should make a tradeoff between its accuracy＂ and complexity and try to keep its former while decreasing the latter. A new model of this kind based on Markov model and directed graph is put forward in this paper. It is designed to have the same accuracy as Markov model but much lower complexity with great flexibleness.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第3期142-144,162,共4页
Computer Applications and Software
