Abstract:  Aiming at the traditional support vector regression machine on the mechanism can’t solute dynamic
time-varying signal pattern classification, proposes a process support vector regression time series prediction
model, and the vortex search algorithm for support vector machine parameter optimization. Using air quality data
set of UCI(University of California Irvine)machine learning repository and belgium solar influences data analysis
center sunspot activities data for simulation. The simulation results show that the prediction results of the
prediction model are better than the particle swarm optimization process support vector regression and support
vector regression, the time series prediction model has well predictive ability.

Key words: process support vector machine, vortex search, parameter optimization