To deal with the problem of detecting the cobalt ion concentration in the purification process of zinc
hydrometallurgy, a cobalt ion concentration combination prediction model based on intelligent fusion strategy is proposed.
In order to improve the prediction accuracy, two online support vector sub-models are established by taking into account the
impact of different kernel functions on the prediction performance. And an improved particle swarm optimization algorithm
is applied to the optimization of parameters in these models. Then the combination prediction model is established through
the intelligent fusion strategy of entropy method. Simulation results show that, the combination model has good prediction
performance which satisfies the error request of the cobalt ion concentration in the zinc sulfate solution purification process.