Foreign Exchange Prediction

Financial markets as the origins of wealth have always been attractive for so many people. In a financial market, e. g. in Foreign Exchange (FOREX) market, the amount of fund and also the correct decision-making are of great importance. With the advances in economic management methods, the latter is now paid more attention. In this way there are successful attempts which provide the investors with powerful decision support systems designed for the market. These are some algorithms which use fundamental analysis, technical analysis or AI-based algorithms to predict the market. In nearly all of these algorithms the target is to provide one step or multi step ahead predictions according to which buy and sell signals will be generated. The recent software, especially those use artificial intelligence and soft computing methods, really challenge some traditional theories like efficient market theory which claim that a financial market is not predictable.

Data in FOREX market is generated in sequence measured typically at successive times. It means that this is a time series problem, a term from statistics and signal processing. These time series are also noisy and nonstationary. All of these made traditional statistical methods inefficient and weak for the prediction task.

Alternatively, in the last decade a lot of attention has been paid to neural networks to solve the problem. Artificial Neural Networks (ANNs) are computational models which are basically inspired by biology, i.e. the human brain. ANNs are usually comprised of densely interconnected, simple, adaptive processing units, which are called neurons. Theses networks may be trained to learn the patterns exist in experimental data, so they may be used for time series prediction problem. Along with the method of artificial neural network, Soft Computing introduces some other powerful methods as fuzzy inference system and genetic algorithm. All these methods provide human like expertise which results in a profitable investment if designed correctly.

Here, at NDL in SRRF, we have used our algorithms and strategies to solve the problem of exchange rate forecasting in FOREX market. Backtesting our AI-based algorithms in FOREX market, we got very good results.