基于动态分类的Markov用户浏览预测模型

Markov User Navigation Prediction Model Based on Dynamic Sorting

针对多Markov链用户浏览预测模型分类算法的时间复杂度过高问题,提出一种基于动态分类的Markov用户浏览预测模型。该模型通过学习提取用户浏览特征,利用这些特征对用户浏览路径进行分类,实现预测并动态更新用户浏览特征。实验结果表明,该模型可明显降低用户浏览路径预测的时间,并得到较为准确的预测结果。

Aiming at the algorithm of Multi-Markov model has higher time complexity, this paper proposes a new approach to model user navigation sequences based on dynamic sorting Markov model. This model gets users navigation characters and uses the characters to sort users and predict users＇ navigation pattern. In particular, this model can shorten the time of prediction obviously and the result is more accurate in prediction.