9. Research on Distributed Principal Component Analysis Algorithm
Project Description: Principal Component Analysis (PCA) is the basis of many data analysis, array processing, and machine learning methods. In extremely large applications involving data arrays, especially in distributed data collection systems, distributed PCA algorithms can use local communication and network connections to overcome the need to locally communicate and access the entire array. The key feature of distributed PCA algorithms is that they ignore the traditional view that the first step of calculation is mainly the formation of the sample covariance. This project is an investigation of the performance of different distributed PCA methods on different data sets, their performance, and their application in distributed data collection systems.