A study on the application of PCA and PSO-SVM to the classification of grape wines
XU Xiao-hua;QUAN Xiao-song;ZHANG Zi-feng;HU Xiao-fei;School of Information Science and Technology,Zhaotong University;School of Mathematics and Statistics,Zhaotong University;
As the data of physical and chemical components of grape winesare characterized by redundancy,this paper proposes a model based on PCA and PSO-SVM for the classification of grape wines. First,it analyzes the principal physical and chemical components of grape wines,and the major influencing factors in order to reduce input dimensions; second,the particle swarm optimization is adopted to find out the best parameters of the support vector machine which is used to complete the study of training samples and forecast the classification of test samples. The result indicates that this model has higher precision than other models and is of certain applicability.