Prediction marketing risk of industrial brand based on optimized support vector machine

Prediction marketing risk of industrial brand is very important which can provide a decision support for top manager. In this paper, we presented an optimized support vector machine (OSVM) to predict marketing risk of industrial brand. Support vector machine (SVM) is a novel statistics learning method, which predict samples using a risk function based on structural risk minimization principle. However, SVM need to determine several parameters, which has a certain influence on classification precision and generalization ability. Paritcle swarm optimization (PSO) is a new optimize method, but it is easy to fall into local optimization and have the low speed of convergence in the late. Generalized pattern search (GPS) have excellent local search ability, but its search result mostly depends on the initial location. To address the issue of SVM, combining the advantages and disadvantages of PSO and GPS, we presented an optimized support vector machine. By incorporating with the GPS, the global searching ability of the PSO is enchanced, and the local searching ability of the PSO is holded, which can optimize the parameters of support vector machine. We apply OSVM to predicte marketing risk of industrial brand. Experimental results show that the prediction accuracy improved by the proposed method.