<![CDATA[The tourism sector becomes one of the pillars in the Indonesian economy. As Bali has been contributing for more than 40 percent of international tourist arrivals in Indonesia. Predicting tourism demand are very important for the government and industry, as predicting the basis for effective policy planning. Support Vector Regression (SVR) is prediction method that has the ability to handle large-scale data in the training phase and it can to recognize patterns of time series data. The predicted result will be good if the value of the important parameters of the SVR can be determined correctly by optimization. One of optimization methods is Genetic Algorithm (GA). GA will be optimizing parameter of SVR to get the right value of SVR parameter to getting better predictions. The test shows the value of MAPE obtained is 2,513% with best parameters those are range of lamda 1 - 10, range of complexity 1 - 100, range of epsilon 0,00001 - 0,001, range of gamma 0,00001 - 0,001, range of sigma 0,01 - 3,5, Iteration of SVR 1250, generation of GA 90, population 70, crossover rate 0,6, mutation rate 0,4, features 2 and prediction period 1 month. Based on the test results, GA-SVR method on the data of foreign tourist arrivals to Bali is appropriate for short-term prediction.]]>
Copyrights © 2018
