Comparison of forex market forecasting tools based on Evolino ensemble and technical analysis indicators

Abstract

Financial markets are an important mechanism for allocating funds to the economy. Traders in finance markets use different strategies to increase their probability of success, and artificial intelligence is already often integrated into the investor support system. The purpose of this article is to compare the possibilities of different trading strategies to detect and predict exchange rate changes. Our model, based on an Evolino ensemble, provides two histograms based on high and low data. Probability estimation, the rejection of unlikely values, is the basis of these strategies, in which two known indicators are compared with strategies based on an Evolino ensemble prediction. Bollinger bands and Ichimoku Kinko Hyo indicators were selected because their lines determine the extreme points of fluctuation regarding exchange rates. Our findings indicate that high and low distributions received by an Evolino ensemble allow the investor to increase the probability of success and can be successfully used to robotize trading in the currency market or to develop new fintech services for investors.