E-Commerce, E-Business and E-Service

164

SVR model parameters, in order to improve the

prediction accuracy and hit ratio of converter

steelmaking endpoint by support vector regres-

sion machine.

2. Support vector regression machine prediction

model (AF-SVR) based on artificial fish swarm

algorithm and put forward here, puts the esti-

mation sample deviation sum minimized as the

optimality principle, by using the strong paral-

lel search performance and the feature of being

insensitive to initial value of artificial fish swarm

algorithm, enhances the performance of AF-

SVR model for converter steelmaking endpoint

prediction in nonlinear processing and parallel

search, etc., and it’s in favour of meeting predic-

tion precision better and faster.

3. In the contrast analysis of experimenta ...