Comparison of Accuracy and Predictive Results of ES-Holt Method and ANN-Back Propagation Method: Case Study of Number of Tourists

Abstract

The COVID-19 pandemic has weakened many sectors, one of which is tourism. Almost all priority tourism locations of the government experience a significant decrease in the number of foreign and domestic visitors. This also impacts the level of welfare of the community around the tourism area. This study is conducted to compare the Exponential Smoothing (ES) Holt and Artificial Neural Network (ANN) Back Propagation methods. That aims to see how the effectiveness and accuracy of both methods to be applied to the predictive process of the number of tourist visits in North Maluku Province using the G-MFS-based MATLAB application. The data on the number of tourists were taken from the Central Statistics Agency of North Maluku Province in 2008-2019. The accuracy parameter indicates that the error for the Holt method has a MAPE value of 9.2976. As for the Back Propagation method, the error is smaller, with a MAPE value of 1.2585. From the results, in 2020 the number of tourists visiting North Maluku should have been around 13,868 people per month. This result can at least be compared to the number of real visitors in 2020 during the COVID-19 outbreak. There is a very high difference between what was predicted and what happened in the field. From the results of this research, it is expected that tourism-related parties or institutions can conduct further research and studies so that they can know the policies to be taken in order to increase the number of tourists visiting North Maluku Province.