Prediction Of Hainan Tourism Volume Based On Baidu Index

It is a challenging task to predict the tourism volume accurately.Research shows that online data(such as Baidu index)is a new source of data that can be used to predict tourism volume.This paper introduces a method to predict tourism volume by using Baidu index.Machine learning technology can be used to further improve prediction accuracy of Baidu index for forecasting tourism volume.Firstly,this paper extracts Hainan tourism volume data and related Baidu index data,constructs the internet search index according to Baidu index data,and verifies the co-integration relationship and Granger causality between the internet search index and Hainan tourism volume.Secondly,the independent variables of the prediction model are divided into time series and “time series+ internet search index”,and machine learning is used to predict the tourism volume.Finally,the prediction effect of different models is evaluated.The experimental results show that compared with the ARIMA model,the forecasting accuracy of the kernel extreme learning machine(KELM)model with independent variable “time series + internet search index” is higher.