Optimization and Application of Support Vector Machine Based on SVM Algorithm Parameters

The hospital customer classification is very important for the integration and allocation of hospital resources, can greatly enhance the market competitiveness of the hospital. Support Vector Machine (SVM) is an approach to solve classification problem by using optimization method. Selecting different kernel parameters can construct different classifiers, meanwhile parameters decide their learning and generalization ability. In order to solve the limitation of selecting parameters by experience, so the particle swarm (PSO) pattern search algorithm is proposed to search optimal parameters and take them into the practice of hospital customer classification; The PSO mode search algorithm is to combine the advantages of PSO algorithm and pattern search algorithm, PSO mode search algorithm has strong global search capability and good advantage of local convergence. The result of experiment shows that this method is not only efficient, but also to search the optimal parameters achieving a high accuracy, which is an effective method of SVM parameter optimization.