PREDIKSI NILAI TUKAR MATA UANG DALAM SISTEM FOREX TRADING MENGGUNAKAN ALGORITMA GENETIKA; PREDICTION OF CURRENCY EXCHANGE RATE IN FOREX TRADING SYSTEM USING GENETIC ALGORITHM

SEPTIAWAN, FAUZI YUDHI, Afiahayati

2016 | Skripsi | FMIPA

In the foreign currency exchange (FOREX) system, there are always fluctuating price changes. Many people attempt to predict currency movements on the FOREX to be able to take appropriate decisions. Prediction method using a genetic algorithm with time-series data is one method that can be used in prediction problems. The genetic algorithm is an algorithm which inspired by the natural evolution. This study used genetic algorithm which begin by generate a population, perform roulette wheel selection method, followed by the whole process aritmethic crossover and random mutation. The results of the population will be selected by the method of steady-state updates. The calculation of the fitness evaluation was conducted using multiple linear regression with the number of variables that can be determined. Level prediction accuracy was measured by using the value of MAPD (Mean Absolute Percentage Deviation) and MSE (Mean Squared Error). This testing process use time-series data for each type of currency data which we want to predict for the span period of 1 year from January 2015 to December 2015. The tests were conducted on four types of major currency pairs in FOREX system that is USDJPY, USDCHF, GBPUSD and EURUSD. MAPD and MSE value which obtained on prediction from January to February 2016 for the USDJPY currency pair at 0,6930% and 1,1186533, USDCHF at 0,7425% and 0,0000734709051153, GBPUSD at 0,7190% and 0,0001600516436978 and for EURUSD at 0,4601% and 0.0001180997.

Kata Kunci : forecasting, FOREX, genetic algorithm, time-series prediction