Neural Network Model Based on Data Preprocessing Technique for Foreign Tourists Prediction

There are various ways done by the government to increase regional income. One of the sectors to increase regional income is tourism sector. The uncertain arrival of foreign tourists makes it difficult for government to predict the number of foreign tourists. The prediction of foreign tourists is very important in assisting decision-making related to regional income. The prediction accuracy using a linear model has limitations in dealing with non-linear data. Thus, a reliable time series prediction model, especially in the field of tourism, is needed. This research proposes soft computing model that is neural network model based on data preprocessing technique for prediction of foreign tourists in Central Java Province, Indonesia. The data of this study are taken from the Department of Youth, Sports and Tourism of Central Java Province, to evaluate the proposed model. The results of this study indicate that neural network model has a better prediction performance in predicting foreign tourists. The results of this study prove that neural network model with NN (2-4-1) has better performance compared to linear regression (trend), moving average, single exponential smoothing,double exponential smoothing, triple exponential smoothing, and ARIMA models.