离群模糊切换回归模型研究

Research on Fuzzy Switching Regression Models with Outliers

模糊切换回归模型被广泛地应用于经济领域和数据挖掘技术中。本文在传统的模糊切换回归模型的基础上,通过为每一个数据样本分配一个动态的权值,得到了一个新的离群模糊切换回归模型,提出了FCWSRM(FuzzyCWeightedSwitchingRegressionModel)算法,最终得到的权重是数据分布的一种代表,将被用来发现数据集中的离群点。实验表明,离群模糊切换回归模型比传统的模糊切换回归模型具有更强的鲁棒性,并且能通过权值发现数据集中的离群点。

Switching regression models have been extensively used in economics and data mining in databases.Based on the classical fuzzy C regression model,by assigning a weight to each datum,this paper obtains a new algorithm of fuzzy C weighted switching regression model(FCWSRM) to find the outliers with fuzzy C regression model.When FCWSRM finally converges,the weight of each datum is a representation of the datum and is used as the criterion to find the outliers in the dataset.Two numeric examples illustrate the effectiveness of FCWSRM.