Study Of MEWMA Control Charts For High-dimensional Data Streams With Noise

In recent years,the monitoring of high-dimensional data flow has always been a hot issue,and the monitoring methods of high-dimensional data have emerged in an endless stream.However,when the sample size is much smaller than the sample dimension,many monitoring methods are less efficient and prone to false positives.In view of this situation,some researchers propose to reduce dimension of the original sample space first,and then monitor the processed data combined with the control chart.Traditional dimensionality reduction methods include principal component analysis(PCA)and linear discriminant analysis(LDA).Although this method can effectively extract most of the information from the data,there is still a certain degree of information loss,which reduces the accuracy of the monitoring results.Notably,due to the complexity of calculation,these dimensionality reduction methods have the problem about slow speed.In recent years,random projection method is widely used in dimensionality reduction for high-dimensional data.Based on the randomness of the selection of projection matrix,this method effectively reduces the computational complexity and improves the operation speed.Similar to the traditional dimensionality reduction method,there is still a certain amount of information loss.In addition,for high-dimensional data streams with noise,PCA and Gaussian random projection matrix can not give full play to their advantages,and the performance of dimensionality reduction is poor.Therefore,this paper mainly focuses on the following two aspects:(1)For the high-dimensional data stream with noise,firstly,the sparse random projection matrix is used to reduce the dimension of the original sample space to 6)different subspaces,then the different subspaces are monitored combined with the MEWMA control chart,and finally integrated;(2)Based on the method proposed in this paper,the sparse random projection matrix is used to reduce the dimension of the patient data of arrhythmia with noise,and then the MEWMA control chart is used to monitor the data.Practice shows that the monitoring scheme in this paper can well monitor the patients with arrhythmia and give early warning...