基于改进的Wap算法的Web序列模式的研究

基于不完备决策表的属性约简定义有多种,现研究基于知识粒度的属性约简。研究发现,差别矩阵是一种较好的设计属性约简算法的方法。为此,定义了一种粒度差别矩阵和基于该差别矩阵的属性约简,并证明了该差别矩阵的属性约简定义与基于知识粒度的属性约简定义等价。在此基础上,设计了一个新的基于信息量的不完备决策表的属性约简算法,其时间复杂度得以降低。

Research on the Web Sequence Pattern Based on the Improved Wap Algorithm

There are many attribute reduction definitions.We researched attribute reduction based on knowledge granuation in incomplete decision table.The discernibility matrix method is a good way to design attribute reduction algorithm.So we gave the definition of discernibility matrix of granulation and the corresponding definition of attribute reduction.At the same time,we proved that the definition is the same as the definition about attribute reduction based on knowledge granulation in incomplete decision table.On this condition,we used the above discernibility matrix of granulation to design an efficient algorithm of attribute reduction based on knowledge granulation in incomplete.Its time complexity is reduced.