Application of Least-Squares Support Vector Regression with PSO for CPU Performance Forecasting

The success of CPU performance prediction will make many benefits. This study adopts the least-squares support vector regression (LS-SVR) with particle swarm optimization (PSO) algorithm to improver accuracy of CPU performance prediction. LS-SVR with PSO, support vector regression (SVR) with PSO, general regression neural network (GRNN), radial basis neural network (RBNN), and linear regression are employed for CPU performance prediction. Empirical results indicate that the LS-SVR (Linear kernel) with PSO has better performance in terms of forecasting accuracy than the other methods. Therefore, the LS-SVR (Linear kernel) with PSO model can efficiently provide credible CPU performance estimated value.

[1]  J. Mercer Functions of positive and negative type, and their connection with the theory of integral equations , 1909 .

[2]  F. Bellotti,et al.  Instruction level analytic prediction of parallel CPU architecture performance , 1997, Proceedings Intelligent Information Systems. IIS'97.

[3]  J. Mercer Functions of Positive and Negative Type, and their Connection with the Theory of Integral Equations , 1909 .

[4]  James Kennedy,et al.  The particle swarm: social adaptation of knowledge , 1997, Proceedings of 1997 IEEE International Conference on Evolutionary Computation (ICEC '97).

[5]  Yifei Wang,et al.  A normal least squares support vector machine (NLS-SVM) and its learning algorithm , 2009, Neurocomputing.

[6]  David W. Aha,et al.  Instance‐based prediction of real‐valued attributes , 1989, Comput. Intell..

[7]  G. G. Stokes "J." , 1890, The New Yale Book of Quotations.

[8]  Gang Xu,et al.  Particle Swarm Optimization-based LS-SVM for Building Cooling Load Prediction , 2010, J. Comput..

[9]  Jeanine Cook,et al.  Ultra-Fast CPU Performance Prediction: Extending the Monte Carlo Approach , 2006, 2006 18th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD'06).