Application of the Machine Learning Method for Predicting International Tourists in West Java Indonesia Using the Averege-Based Fuzzy Time Series Model

Abstract

The purpose of this study is to propose whether an average-based fuzzy time series model is appropriate for use in predicting the number of foreign tourists coming to West Java, Indonesia. Machine learning is a branch of artificial intelligence where machines are designed to learn on their own without human direction. One of the machine learning methods used by data science is for prediction processes, such as predicting the number of tourists. Tourism is one of the economic sectors that has a direct impact on the community's economy. Based on data from the Badan Pusat Statistik (BPS), the number of tourists coming to West Java Indonesia fluctuates, meaning that the number can increase and decrease every month and year. Changes in the number of tourists that fluctuate are one of the problems that have an impact on tourism actors. Therefore, the solution given to answer this problem is that an appropriate model is needed to predict the number of tourists visiting West Java. The contribution of this research is to help related parties in predicting the number of foreign tourists so that it can be used as one to make policies related to tourism preparation and planning efforts in West Java, Indonesia.  The method used in this research is a case study approach, where the case study is taken from data on foreign tourists visiting West Java from 2017 to 2020. For the prediction process, the method used is the fuzzy time series method and the average length-based algorithm as the determinant of the interval length. Effective interval length can affect prediction results with a higher level of accuracy. Based on the prediction test results, the Mean Absolute Percentage Error (MAPE) value is 14.71%. These results indicate that the fuzzy time series model based on the average interval length is good for prediction.