Abstract: Web mining is baseding on excavating of text stream,and owing to the fact that the characteristic of sample vector often has several ten thousand, the operation speed of classification algorithm directly influences his reality applying. Putting forward to the fuzzy support vector machine classification algorithm based on T-S model, the superiority of algorithm embodies several aspects in the following, first, utilizing its language information efficiently. Ssecond, the difficulty of determining optimum solution via two times planning for mass data is solved because two times planning optimum solution could be determined through only partial samples; finally, it could be known through the algorithm that it could be executed on parallel operation running on computers and this will improve the speed of the algorithm.

Key words: Web mining, fuzzy support vector machine, T-S fuzzy model

摘要： WEB挖掘是基于文本流的挖掘，由于样本向量的特征往往有几万个，分类算法的运算速度直接影响其实际应用。提出了基于T-S模型的模糊支持向量机分类算法，算法的优势体现在下面几个方面，第一，充分利用了语言信息。第二，由于只需通过局部样本求解二次规划最优解，因此，解决了海量数据的二次规划求最优解的困难。第三，从算法中可看出，在计算机上其算法可实行并行运算，这样提高了算法的运行速度。

关键词: WEB挖掘, 模糊支持向量机, T-S模糊模型