Forecasting of Foreign Tourists' Arrivals in Bangladesh: A Neural Network Approach

Toukir Ahmed, Nripon Mollick, Khairun Nahar
2019 Zenodo  
Forecasting of foreign tourists' arrivals in any destination country is very important for the tourism industry of that country. It has direct impact on social, cultural, educational and economic sectors of most countries. In this paper a feed-forward neural network model was used to forecast foreign tourists' arrivals in Bangladesh. Some other models like naive, simple moving average, single exponential smoothing and multiple regression model were used for comparison with neural network model.
more » ... One output variable and seven input variables were used. The output of the neural network model represented the foreign tourists' arrivals in Bangladesh. Data from various sources for the previous thirty years were used. The estimated foreign tourists' arrivals were compared with the actual officially published foreign tourists' arrivals. Empirical results showed that neural network model outperformed other forecasting models. For the purpose of forecasting the next ten year's foreign tourists' arrivals by neural network model, two other methods named double moving average method and trend forecasting method were used to predict the variables. Finally next ten years foreign tourists' arrivals were forecasted by the neural network model using the predicted values of the variables from trend forecasting method as it showed higher accuracy than double moving average method.
doi:10.5281/zenodo.4581094 fatcat:teoc36cd4zbtphghhsskihruzu