In order to further improve the prediction accuracy of cigarette consumption amount, a prediction model of cigarette consumption amount (PCA-SVM) combined with principal component analysis (PCA) and support vector machine (SVM) is proposed. Firstly, adopt PCA to preprocess the influence factors of cigarette consumption amount, eliminate redundant information among factors, and lower the input dimension of SVM; then select training set of cigarette consumption amount of SVM according to the input dimension so as to build prediction model of cigarette consumption amount, and use genetic algorithm to optimize SVM parameters; finally, adopt specific dada of cigarette consumption amount to carry out simulation experiment so as to inspect the validity of PCA-SVM.
展开▼