Improving regression models using simulated annealing for stock market speculation

Abstract

The Forex aims at exchanging the so-called convertible currencies from one specific currency to another worldwide. The currency exchange rates can be increased or reduced according to time, between various participants (particular investors, central banks and enterprises). The main pillar of the Forex market is the temporal prediction of the currency exchange rate; it must be well-forecasted to invest in currencies and to maximize profits which will make the speculation more flexible. In the literature, many papers talk about the combination of two methods to improve the prediction of currency exchange. In this paper we propose a hybrid model which is combining both the regression algorithm and the simulated annealing algorithm in order to predict the daily exchange rates of the USD/EUR pair. Finally, the experiments validate that the Hybrid model of the regression algorithm and the simulated annealing algorithm can be beneficial for the prediction of exchange rates.