Clustering is a popular strategy to improve the performance of the randomized block Kaczmarz methods, but it is unavailable for large-scale linear systems due to the substantial complexity associated with clustering high-dimensional data. However, for high-dimensional datasets, the clustering with dimensionality reduction could overcome the aforesaid drawback while achieving comparable clustering results. The Achlioptas random projection, as a powerful dimensionality reduction method, projects high-dimensional data into a low-dimensional space and preserves distance relationships between data points. In this paper, we propose a fast randomized block residual steepest descent method, built upon the Achlioptas random projection and the Gaussian mixture model, for solving large sparse, overdetermined linear systems. The theoretical analysis of which is also established. Numerical experiments are performed to illustrate the effectiveness of the proposed method compared with some existing ones, especially in computing time.
扫码关注我们
求助内容:
应助结果提醒方式:
