基于剪枝技术的自适应PPM预测模型

An Adaptive PPM Prediction Model Based on Pruning Technique

建立有效的用户浏览预测模型,对用户的浏览行为进行准确的预测,是Web预取的关键。标准PPM预测模型由于存在存储复杂度高、执行效率低等缺点,影响了其推广和应用。文章基于剪枝技术,依据Zipf法则及Web对象访问特征对标准PPM预测模型进行预先剪枝和后剪枝,构造出一种自适应PPM预测模型。实验表明,该模型不仅能动态预测用户的Web浏览特征,而且在预测准确率和存储复杂度方面都有一定程度的提高。

The key issue of Web prefetehing is to establish an effective user browsing prediction model,which can be used to make precision prediction of user browsing actions.The high space complexity and low efficiency of the standard PPM prediction model affect its application in Web prefetching.This paper makes use ot pruning technique and proposes a new approach to modeling user navigation sequences based on Zipf＇s law and Web access characteristic.The experiments have shown that this model not only can be used to make predictions of the Web access characteristic dynamically,but also has lower space complexity and more prediction accuracy.