BSP-Based Support Vector Regression Machine Parallel Framework

Abstract

In this paper, we investigate the distributed parallel Support Vector Machine training strategy, and then propose a BSP-Based Support Vector Regression Machine Parallel Framework which can implement the most of distributed Support Vector Regression Machine algorithms. The major difference in these algorithms is the network topology among distributed nodes. Therefore, we adopt the Bulk Synchronous Parallel model to solve the strongly connected graph problem in exchanging support vectors among distributed nodes. In addition, we introduce the dynamic algorithms which can change the strongly connected graph among SVR distributed nodes in every BSP’s super-step. The performance of this framework has been analyzed and evaluated with KDD99 data and four DPSVR algorithms on the high-performance computer. The results prove that the framework can implement the most of distributed SVR algorithms and keep the performance of original algorithms.

Keywords

• parallel computing; bulk synchronous parallel; support vector regression machine (SVR); regression prediction.