I. Introduction
The rapid development of computers in the 21st century has brought forth a numerous amount of data, making it easier and more practical to collect data. Big data and high-dimension play a unique role in various fields such as biochemical science, engineering, finance, and social science. This implies that, when dealing with space-time issues, the machine learning function needs to handle numerous features, images, and other targets extracted from documents. Typical and key features of big data include the large amount and high dimension of samples; sparse data is another feature related to big data [1], [2]. The collection of big data at different positions may lead to the error of measurement or change of experimental platform. In the optimized evolutionary algorithm (EA), the high-dimensional data would deteriorate the performance, which would increase the calculation costs, uncertainty of the algorithm, false correlation of data, as well as accumulation of more noises.