FORECASTING THE NUMBER OF FOREIGN TOURISTS WHO VISIT TO EAST JAVA USING MONTE CARLO METHOD

Indonesia has many islands and there are beautiful inland areas, interesting historical and cultural ruins, beaches, mountains, and more. Especially in the tourism sector is one of the largest industries that are very influential and grow very fast. The advancement of the tourism industry in a region is very dependent on the number of tourists who come both domestic and foreign tourists. The large number of foreign tourists that come push and accelerate economic growth. So that directly leads to an increase in demand for goods and services. To meet the needs and demands of tourists, it is necessary to predict the number of visits of foreign tourists. One method that can be used in forecasting is Monte Carlo. From the results of Monte Carlo research can work well, From the stage of the prediction system implementation that has been built using the initial parameter 12 months 100x simulation and delta-t = 0.001, then get sigma = 52.2650054, Mu = -0.0398. And the simulation is more accurate in predicting the number of foreign tourist visits in East Java, which has a small error value. To get a smaller error value is by reducing or delta-t value.


INTRODUCTION
Travelers are people who travel to vacation, seek treatment, do business, exercise and study and visit beautiful places or a particular country.The word tourist comes from Sanskrit, from the origin of the word "tourist" which means the trip is added with the suffix "wan" which means people who travel.In English, the person who travels is called a traveler.Whereas people who travel for tourist destinations are called Tourist.Foreign tourists are foreigners who travel, who come to another country which is not the country where they usually live.Foreign tourists are also called foreign tourists or abbreviated as foreign tourists.
In the tourism sector is one of the largest industries that are very influential and grow very fast.The development of tourism is encouraging and accelerating economic growth.The advancement of the tourism industry in a region is very dependent on the number of tourists who come both domestic and foreign tourists.so that it directly raises the demand for goods and services.In an effort to meet tourist demand, investment in transportation and communication, hospitality and other accommodation, the handicraft industry and the consumer product industry, service industries, restaurant restaurants and others are needed.So from that research is needed to help the above problems in predicting foreign tourist visits.
The Monte Carlo method of the term in the simulation was introduced by compte de buffon in 1997.It is a computational algorithm to simulate various behaviors in the systems of physics and mathematics.This method is used to evaluate integrals, calculus, and other numerical methods.This method is proven to be efficient in solving differential terrain radians, so this method is used in global illumination calculations that produce photorealistic images of three-dimensional models, which are applied in video games, architecture, design, computer-generated films, special effects in film, business, economics, and other fields.Based on the description above, this study ISSN: 2528-0260 P.479-488 aims to predict foreign tourist visits using the Monte Carlo simulation method.SYSTEM ANALYSIS AND DESIGN.
At this stage of system analysis, which has the task of conducting a feasibility study, analyzing system requirements in the form of processing tourist data, forecasting calculations of tourist visits and forecasting report results.The data used in this final project is data on the number of visits of foreign tourists visiting East Java.

Figure 1. Tourist Visit System Flowchart
In the flowchart image above can be explained to start a computing system that is by logging in first by the user / admin, then inputting the data to be predicted.The next stage of Monte Carlo calculations input many months and the number of simulations then calculates the delta-t ( t) variable, deviation (s), sigma ( ), Mu (µ).So as to produce forecasting the number of visits.Then calculate errors using MSE and MAPE.

Test Predictions
First try: The initial value of 12 months by simulating as much as 50x gets sigma value = 0.00021062, Mu = 0.00000053, delta-t = 0.08333, deviation = 0.0000608, and gets a random value = 0.920578.The following forecasting data can be seen in table The following is the result of the first trial screenshot in Figure 8:

Experiment Analysis
From the 5x results the experiment uses the initial value of the 12-month parameter by simulating 50x, 100x, 150x, 200x, 250x and getting a sigma value = 0.00021062, deviation = 0.0000608, delta-t = 0.083333, Mu = 0.00000053 and produces a different random value.

Test Results
From several experimental results obtained, it can be concluded that using 12-month input parameter values and 250x simulations obtain an overall predictive value of 220571.9145.Based on the comparison of the value of MSE = 0.02617 with MAPE = 0.00087, it can be concluded that using the MAPE method to find a better error value in this study.To overcome the increasing number of visitors in the coming year, socialization is needed for the needs of goods and services.

CONCLUSION 3.1 Conclusions
The conclusions obtained in this study are: (1) The implementation of the system results in a prediction of the number of visits of foreign tourists visiting East Java using the Monte Carlo method to reference the accuracy of the forecasting process.
(2) Implementation of the system using the Monte Carlo method can work well to predict the number of visits of foreign tourists.
(3) From the 5x results the experiment uses the initial value of the 12-month parameter by performing a simulation of 50x, 100x, 150x, 200x, 250x.And get a sigma value = 0.00021062, deviation = 0.0000608, delta-t = 0.083333, Mu = 0.00000053, and produce different random values.
(4) From several experimental results obtained, it can be concluded that using 12-month input parameter values and 250x simulations obtain an overall predictive value of 220571.9145.
Based on the comparison of the value of MSE = 0.02617 with MAPE = 0.00087, it can be concluded that using the MAPE method to find a better error value in this study.

Figure 8 .
Figure 8. Results of the First Trial Program

Table 1 .
Forecasting Result in 2017 on the first try 6.1 and figure 6.1 obtaining MSE = 1.09008 and MAPE = 0.005601.

Table 2 .
Test Results from several experiments