基于时间序列修正算法的我国入境旅游人数预测

Time Series Prediction Correction Algorithm Based on Number of Inbound Tourists in China

为了对我国入境旅游游客量进行准确预测,提出一种将ARIMA模型与RBF神经网络相结合的算法。以我国2009年1月到2014年4月我国入境旅游游客量月度数据为研究对象,利用该模型对我国入境旅游游客量进行初步预测,计算残差,再利用RBF神经网络对残差进行拟合预测,并对ARIMA预测结果进行修正。结果表明:利用RBF神经网络对ARIMA模型进行修正,将线性拟合算法和非线性拟合算法结合起来用于我国入境旅游游客量预测是一种较可靠的算法。

In order to carry on Chinese inbound tourists quantity forecast, this paper proposed a ARIMA model and RBF neural network algorithm. In 2009 January to 2014 April, taking Chinese inbound tourism tourist amount monthly data as the research object, this paper firstly used the ARIMA prediction of Chinese inbound tourists to compute the residual. Then using the RBF neural network to forecaste the residual and the ARIMA prediction results was correction. The methods used for the Chinese inbound tourism tourist quantity prediction, the results showed that：The ARIMA model was modified by using the RBF neural network, the linear fitting algorithm and nonlinear fitting algorithm were combined to predict Chinese inbound tourists quantity which was a more reliable algorithm.