•Soft Computing (SC) hybrids for forecasting FOREX rate were proposed to obtain predictions that are more accurate.•A comprehensive review of 82 articles published during the 1998–2017 presented•All ...hybrids outperformed their constituents in terms of accuracy.•ANN-based hybrids turned out to be more pervasive and more powerful.•Both Evolutionary Computation and Fuzzy logic based hybrids do also contain some neural networks as a predominant constituent.•Any future research might necessarily contain either traditional or sophisticated neural networks or Support Vector Machine.•Critique of all families of soft computing hybrids is presented along with future directions
Foreign exchange rate prediction is an important problem in finance and it attracts many researchers owing to its complex nature and practical applications. Even though this problem is well studied using various statistical and machine learning techniques in stand-alone mode, various soft computing hybrids were also proposed to solve this problem with the aim of obtaining more accurate predictions during 1998–2017. This paper presents a comprehensive review of 82 such soft computing hybrids found in the literature. Almost all authors in this area demonstrated that their proposed hybrids outperformed the stand-alone statistical and intelligent techniques in terms of accuracy. It is conspicuous from the review that artificial neural network based hybrids turned out to be more prevalent, more pervasive and more powerful. This observation is corroborated by the fact that both evolutionary computation based hybrids as well as fuzzy logic based hybrids also contained some architecture of neural networks as a predominant constituent. The review concludes with a set of insightful remarks and future directions that are very much useful to budding researchers and practitioners alike.
Financial market predictions represent a complex problem. Most prediction systems work with the term time window, which is represented by exchange rate values of a real financial commodity. Such ...values (time window) provide the base for prediction of future values. Real situations, however, prove that prediction of only a single time-series trend is insufficient. This article aims at suggesting a novelty and unconventional approach based on the use of several neural networks predicting probable courses of a future trend defined in a prediction time window. The basis of the proposed approach is a suitable representation of the training-set input data into the neural networks. It uses selected FFT coefficients as well as robust output indicators based on a histogram of the predicted course of the selected currency pair. At the same time, the given currency pair enters the prediction in a combination with another three mutually interconnected currency pairs. A significant output of the articles is, apart from the proposed methodology, confirmation that the Elliott wave theory is beneficial in the trading environment and provides a substantial profit compared with conventional prediction techniques. That was proved in the performed experimental study.
•It was proved the benefits of the Elliott Wave theory in trading simulation.•It was proved a real financial profit on a sufficiently large test samples.•It was proposed an innovative approach for suitable representation of the input data.•The proposed approach successfully minimizes the pattern offset problem.•The successfulness of the presented trading system achieves 77% on average.
Investing in the stock market and Forex can be lucrative, but it is important to approach it with caution and a clear understanding of the risks involved. Predicting the direction of prices in ...financial markets is a complex task, and there is no guaranteed way to do it. One innovative approach that has been proposed involves using a combination of the kinetic energy formula and indicator signals to predict prices, besides another predictions using deep reinforcement learning (DRL). This approach has led to the development of the Trading Deep Q-Network algorithm (TDQN), which incorporates the kinetic energy of stocks/currencies as a condition rule. The proposed approach, TKDQN method, has shown promising results in terms of accuracy and profitability, outperforming previous versions based on several metrics.
Using technical price patterns is one of the well-known techniques for predicting future trends in financial markets. Some of these patterns are profitable under certain conditions and some might be ...non-profitable based on the target market situation and spread. This paper aims to propose a model that works along with the moving average crossover technical pattern. The outputs of the technical price pattern, which are long or short signals, are given as input to the proposed model to predict its profitability. We use a joint model that benefits from two different types of intelligent processing techniques, namely image processing which is applied to candlesticks extracted from price history, and time series analysis which is applied to the numerical features. For the former process, Convolutional Neural Network (CNN) is used and for the latter process, CNN with Long Short-Term Memory (LSTM) is used for the prediction. The proposed model is applied to the data from EUR/USD pairs. The tests were performed for spread values of 0.5, 1, 1.5, and 2. We show that the hybrid model achieves superior results compared to the individual ones, Relative Strength Index (RSI) and Bollinger Bands (BB) technical analysis patterns, as well as two state-of-the-art price prediction models based on CNN-Bidirectional LSTM (BiLSTM) and Phase-State Reconstruction (PSR) with LSTM.
•A hybrid model enhances moving average crossover technical pattern signals.•The profitability of the moving average crossover signal outputs is predicted.•Deep image processing and time series analysis are used to classify the signals.•The hybrid model achieved better results in comparison to state-of-the-art baselines.
Display omitted
•Forecasting on time series data from finance domain (Forex).•Using genetic algorithm for parameter selection and rule combination.•Generating trading rules using technical ...indicators.•Using greedy search heuristic for rule selection and combination.•Applying hybrid evolutionary methods on real life very large data set.
Technical indicators are widely used in Forex and other financial markets which are the building blocks of many trading systems. A trading system is based on technical indicators or pattern-based approaches which produces buy/sell signals to trade in the market. In this paper, a heuristic based trading system on Forex data, which is developed using popular technical indicators is presented. The system grounds on selecting and combining the trading rules based on indicators using heuristic methods. The selection of the trading rules is realized by using Genetic algorithm and a greedy search heuristic. A weighted majority voting method is proposed to combine the technical indicator based trading rules to form a single trading rule. The experiments are conducted on 2 major currency pairs in 3 different time frames where promising results are achieved.
Statistical and multiscaling characteristics of WTI Crude Oil futures prices expressed in US dollar in relation to the most traded currencies as well as to gold futures and to the E-mini S&P500 ...futures prices on 5 min intra-day recordings in the period January 2012–December 2017 are studied. It is shown that in most of the cases the tails of return distributions of the considered financial instruments follow the inverse cubic power law. The only exception is the Russian ruble for which the distribution tail is heavier and scales with the exponent close to 2. From the perspective of multiscaling the analysed time series reveal the multifractal organization with the left-sided asymmetry of the corresponding singularity spectra. Even more, all the considered financial instruments appear to be multifractally cross-correlated with oil, especially on the level of medium-size fluctuations, as the multifractal cross-correlation analysis carried out by means of the multifractal cross-correlation analysis (MFCCA) and detrended cross-correlation coefficient ρq show. The degree of such cross-correlations is however varying among the financial instruments. The strongest ties to the oil characterize currencies of the oil extracting countries. Strength of this multifractal coupling appears to depend also on the oil market trend. In the analysed time period the level of cross-correlations systematically increases during the bear phase on the oil market and it saturates after the trend reversal in 1st half of 2016. The same methodology is also applied to identify possible causal relations between considered observables. Searching for some related asymmetry in the information flow mediating cross-correlations indicates that it was the oil price that led the Russian ruble over the time period here considered rather than vice versa.
•High frequency world oil price fluctuations obey the inverse cubic power-law.•World currencies, gold and S&P500 are multifractally cross-correlated with the oil.•The degree of this cross-correlation depends on the oil market phase.•Russian ruble reveals the strongest correlation and is led by the oil market.•MFCCA allows to detect at what range of fluctuations the cross-correlations dominate.
In the decade passed, considerable affords were made to develop effective trading systems based on different assumptions concerned with the market nature, methods for data processing and uncertainty ...modeling. Such systems are often so sophisticated that they can be applied only by their authors. Another limitation of them is concerned with the focus on the development of a universal single best model. Besides, any model works well only in limited time periods and fails when noticeable changes in the market behavior occur. Then a major revision or the development of a new model is inevitable. Unfortunately, usually this needs too much time. Therefore, in this paper, to avoid the above problems, the simple multi-model approach to the development of trading systems in the Forex market is proposed. It is based on some working hypotheses, which are justified in this paper. The first of them is based on the observation that the Forex is the aggregation of numerous streams (strategies) provided by the broad trades community. Therefore, we can expect that even a very simple model based on the particular trading idea or ideas may catch such a string to be profitable, at least during a small period. If we have developed a set of such simple models optimized for different currency pairs, in each trading period we can use the model providing maximal profit for a certain currency pair. The profitability of the proposed approach is illustrated by the trading results obtained on the symbols EURUSD,GBPUSD, AUDUSD and USDJPY for the timeframes H1 and H4 with the use of the Meta Trader 4 platform.
•Technical analysis indicators and trading rules.•Positive and Negative Overfitting effects.•Algorithmic trading systems and their open access software.•Multi-model based trading strategy and the leader correction method.•Forex market and trading platform.
•Examine asymmetries in the volatility spillover of international currency markets.•Focus on frequency based spillover and the COVID-19 pandemic.•Propose partial quantile coherency network ...approach.•Networks are driven by developed currency markets and by geographical proximity in Europe and Asia.•Dependence structure changes during COVID-19 especially in the long run.
We examine asymmetries in the volatility spillover of international currency markets over the short and long run, with a focus on the COVID-19 pandemic. In doing so, we propose partial quantile coherency network approach. Our results indicate heterogeneous behaviour of currencies’ volatility networks under various market conditions across investment time horizons. The volatility networks are driven by developed currency markets and by geographical proximity in Europe and Asia. We do not find asymmetry in the dependence structures of positive and negative currency volatilities. The dependence structure changes during COVID-19 especially in the long run. Many currencies show disentangled behaviour, which suggests their hedging and diversification potential.