Currency trading in volatile markets: Did neural networks outperform for the EUR/USD during the financial crisis 2007–2009?

Abstract

The motivation for this article is to check whether neural network models have remained a superior method for forecasting the EUR/USD exchange rate during the financial crisis of [2007][2008][2009]. Alternative neural network architectures (Multi-Layer Perceptron (MLP), Recurrent Neural Network and Higher Order Neural Network (HONN)) are benchmarked against a random walk and a traditional ARMA model, and evaluated in terms of statistical accuracy and through a trading simulation on daily data over the period from January 2000 to February 2009, the period from August 2007 to February 2009, providing the out-of-sample testing period. Transaction costs and a confirmation filter devised to reduce false signals and thus also reduce losses and transaction costs were also taken into consideration. It is shown that the HONN structure gives the overall best results on a simple trading simulation; however, for an advanced trading simulation with a confirmation filter, the MLP outperforms all other models on most performance measures. On the whole, the results show that neural networks are still able to produce forecasts that yield a positive return and are superior to those of linear and more traditional models, with respect to both trading performance and statistical accuracy, even under very volatile market conditions.

A review of artificial intelligence quality in forecasting asset prices

“…The results showed a substantial difference in performance, with lower errors in the model with normalized data and a lower MAPE, which supports the findings of previous authors. Dunis et al (2012) and Jasic and Wood (2004), however, have commented that normalization through data transformation is somewhat controversial in the literature and requires further examination in specific studies. Jasic and Wood (2004) argued that the positive effects of normalization may decrease with network adjustment, particularly when the network size and sample increase.…”

“…All numbers are percentages. (Araújo, 2010;Khairalla & Al-Jallad, 2017;Li et al, 2018;Rath et al, 2019;Rout et al, 2014) perform better than the nonhybrid top-five (Dunis et al, 2012;Karathanasopoulos & Osman, 2019;Karathanasopoulos, 2017;Lakshmi & Visalakshmi, 2016;Xu, et al, 2011) with an estimated 0.65% versus 0.87%, respectively. Data normalization proves to be effective in improving predictive power.…”

Design of Experiment Pada Artificial Neural Network Untuk Memprediksi Kurs Tukar Mata Uang Idr/Usd

“…(Ye, 2012) menggunakan ANN untuk memprediksi kurs tukar RMB/USD dan memperoleh hasil forecast yang cukup akurat dan efisien, serta memproyeksi nilai RMB akan meningkat di masa mendatang. (Dunis et al, 2012) menggunakan data harian selama 10 tahun untuk mengevaluasi apakah ANN masih superior dibanding metode lainnya untuk memprediksi EUR/USD, dan hasilnya adalah ANN memberikan hasil yang lebih baik dibanding ARMA dan random walk.…”

A Comparison of Linear and Nonlinear Models in Forecasting Market Risk: The Evidence from Turkish Derivative Exchange

“…Cinko and Avcı (2007) indicate that ANN model is more successful than linear regression model in daily and seasonal forecasting of ISE-100 index. Dunis et al (2012) (2011) compare the performance of traditional ARCH-GARCH models and ANN model and conclude that ANN model has better performance than traditional ARCH-GARCH models. Different from these papers, Mantri etal.…”