摘要
不同数据和不同模型产生不一样的预测结果,也反映了不同的技术范畴,因此,了解不同模型预测的准确性显得尤为重要。为了提高预测的准确性,文章在考虑了时间序列季节性波动的基础上,采用线性模型及非线性模型对中国入境旅游人数、不同目的入境的外国游客数指标数据进行拟合及预测,继而评价两类模型预测结果。结果表明非线性预测模型优于线性预测模型,采用前两年(m=8)作为输入数据进行递归的SVM模型在所有SVM模型中表现良好,对比SVM、MPL和RBF三种模型,RBF模型表现较好。
Different data and different models produce different prediction results and reflect different technical categories.Therefore,it is particularly important to know the accuracy of different model predictions.In order to improve the accuracy of the prediction,this paper uses the linear model and non-linear model to fit and predict the index data of the number of Chinese inbound tourists and the number of foreign tourists for different purposes on the basis of the seasonal fluctuation of time series,and then makes an evaluation on the prediction results of the two models.The results show that the nonlinear prediction model is better than the linear prediction model;The SVM model using the first two years(m=8)as input data for recursion performs very well among all SVM models;compared with SVM,MPL model,RBF model has better performance.
作者
宋鑫
王维国
Song Xin;Wang Weiguo(College of Economics,Dongbei University of Finance&Economics,Dalian Liaoning 116025,China;College of Tourism and Geographic Sciences,Jilin Normal University,Siping Jilin 136000,China)
出处
《统计与决策》
CSSCI
北大核心
2020年第2期5-10,共6页
Statistics & Decision
关键词
回归预测
神经网络
入境旅游
regression prediction
neural network
inbound tourism
