基于 LSTM 和 ARIMA 的组合模型对入境游客人次的预测

叶燕霞

旅游研究 ›› 2018, Vol. 10 ›› Issue (6) : 29-40.

PDF(1265 KB)
PDF(1265 KB)
旅游研究 ›› 2018, Vol. 10 ›› Issue (6) : 29-40.
论文

基于 LSTM 和 ARIMA 的组合模型对入境游客人次的预测

作者信息 +

The Prediction of Inbound Tourists Based on theCombined Model of LSTM and ARIMA

Author information +
文章历史 +

摘要

随着我国经济水平的提高, 旅游业的发展也越来越快, 游客人次的准确预测显得尤其重要。 主要提出了一种基于 LSTM 神经网络和 ARIMA 模型的组合模型对入境游客人次进行预测, 并以香港、 澳门、 上海的入境游客人次为例, 进行了实证研究。 三个城市的入境游客人次的实证结果都表明动态神经网络 LSTM 比静态神经网络 BP 网络更适合预测时间序列,组合模型的预测较准确。 基于 LSTM 与 ARIMA 的组合模型能较准确地预测入境游客人次, 对制定出更合理的旅游资源配置方案, 具有一定的参考价值和实践意义。

Abstract

With the improvement of economy in China, the development of tourism has been fasterand faster, the accurate prediction of visitors is particularly important. A combined model based onLSTM neural network and ARIMA model is proposed to predict the inbound tourists. A case study isconducted by taking inbound tourists from Hong Kong, Macao and Shanghai as examples. Theempirical results of inbound visitors from three cities indicate that the dynamic neural network LSTM ismore suitable for the prediction of time series than the static neural network BP network. Meanwhile,the prediction effect of the combined model is more accurate. The combined model based on LSTM andARIMAis able to predict the quantity of inbound visitors more accurately. There is a certain referencevalue and practical significance for making more rational allocation scheme of tourism resources.

关键词

入境旅游人次 / ARIMA 模型 / LSTM 神经网络 / 组合模型

Key words

inbound tourists / ARIMA model / LSTM neural network / combined model

引用本文

导出引用
叶燕霞. 基于 LSTM 和 ARIMA 的组合模型对入境游客人次的预测. 旅游研究. 2018, 10(6): 29-40
The Prediction of Inbound Tourists Based on theCombined Model of LSTM and ARIMA. Tourism Research. 2018, 10(6): 29-40
PDF(1265 KB)

967

Accesses

0

Citation

Detail

段落导航
相关文章

/