季节调整的PSO-SVR模型及其在旅游客流量预测中的应用——以海南省三亚市为例

Seasonal Adjustment' PSO-SVR Model and It's Application on Tourism Flow Forecast:Taking Sanya as an example

准确的旅游客流量预测对旅游目的地做好事前准备工作至关重要.然而旅游客流量具有明显的非线性和季节性特征,采取季节调整方法对样本数据进行预处理,消除季节性的影响,可以提高客流量预测的准确性.同时SVR(支持向量回归机)是一种良好的机器学习方法,非常适合预测研究,辅以PSO(粒子群算法)选取合适的回归参数可以获得更加精确的预测结果.提出了一种考虑季节影响并通过PSO优化SVR模型的旅游客流量预测模型,并以海南省三亚市为例进行了实证研究.研究结果表明,季节调整的PSO-SVR模型预测精度明显高于SVR、季节调整的SVR和PSO-SVR模型,是进行旅游客流量预测的有效工具.

Accurate tourism flow prediction is crucial to do a good job of preparation for any tourism destinations. Tourist industry has a strong seasonal trend. Taking seasonal adjustment method for preprocessing of sample data to eliminate the seasonal influence can improve the accuracy of tourism flow prediction. At the same time, SVR is a good machine learning method, very suitable for predicting tourism flow, and supplemented by the PSO algorithm to choose appropriate parameters, it can obtain more accurate predicting results. In view of this, this paper study a tourism flow prediction method which consider seasonal influ- ence and through SVR （support vector regression machine） optimized by the PSO （particle swarm optimization） model with Sanya as an example for empirical research. The results indicate that the prediction accuracy of seasonally adjusted PSO-SVR model is significantly higher than the SVR model, seasonally adjusted SVR model, the PSO-SVR model. The seasonally adjusted PSO-SVR model is an effective tool of tourism flow prediction.