Foreigner Visits Estimation Based on Multi Support Vector Machine
Indra Gunawan, Wahyu Sakti Gunawan Irianto, Aji Prasetya Wibawa, Triyanna Widiyaningtyas, Anusua Ghosh, Tinton Dwi Atmaja · 2018
Over time, tourist visit has increased significantly. Data mining has been used to estimate the rate of tourist. Similarly, Support Vector Machine (SVM) has been applied in several studies which proved that it has an accurate estimate value. This study deployed SVM and Multi-Support Vector Machine independently to estimate the tourist visit rate in Indonesia. The results obtained from this study indicate that the data based on the arrival gate point has a fairly low level of correlation between gates and time series data has an influence on processing the number of foreign visits. The average rate of Mean Absolute Persetage Error (MAPE) that resulted from processing data using separately multi SVM was at 0.57 % and integrated multi SVM was at 19.6%.