Comparative analysis of the application of Deep Learning techniques for Forex Rate prediction

Aryal, Saugat; Nadarajah, Dheynoshan; Kasthurirathna, Dharshana; Rupasinghe, Lakmal; Jayawardena, Chandimal

doi:10.1109/icac49085.2019.9103428

Cited by 6 publications

(5 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although there is a trade-off between computation time and prediction performance of LSTM versus Bi-LSTM approaches, the differences in computation time are not significant. Moreover, the proposed three layers Bi-LSTM networks could achieve similar prediction performance results compared to other machine learning and deep learning methods that used deeper network's architecture, as reported in Aryal et al [32] who used LSTM, convolutional neural networks (CNN), and temporal convolution networks (TCN); Qi et al [33] who used RNN, LSTM, Bi-LSTM, and gated recurrent unit (GRU); and Dautel et al [34] who employed FNN, RNN, LSTM, and GRU. Moreover, we could also try to compare the prediction results from this study with other Machine and Deep Learning methods commonly used in the literature, such as naïve Bayes [35], GRU [36], and Random Forest Regressor [37], a popular tree-based algorithm.…”

Section: Performance Resultssupporting

confidence: 58%

On searching the best mode for forex forecasting: bidirectional long short-term memory default mode is not enough

Hansun

Putri

Hugeng

2022

IJ-AI

View full text Add to dashboard Cite

Presently, the Forex market has become the world’s largest financial market with more than US$5 trillion daily volume. Therefore, it attracts many researchers to learn its traded currency pairs characteristics and predict their future values. Here, we propose simple three layers Bidirectional long short-term memory (Bi-LSTM) networks for Forex forecasting with four different merge modes. Moreover, the proposed model is also compared to the conventional long short-term memory (LSTM) networks with the same architecture. Five major Forex currency pairs, namely AUD/USD, EUR/USD, GBP/USD, USD/CHF, and USD/JPY, with more than ten years of historical records are considered in this study. It is revealed from the experimental results that among four available merge modes, the concatenation mode as the default merge mode in Bi-LSTM networks is actually the less preferred mode for Forex forecasting (Root mean square error 0.30685517, mean absolute error 0.27442235, mean absolute percentage error 0.827108%). Moreover, Bi-LSTM average mode gets the highest R2 score that could achieve 89.579%. Therefore, the proposed three layers Bi-LSTM networks could provide a baseline result for developing a good trading strategy in Forex forecasting.

show abstract

Section: Performance Resultssupporting

confidence: 58%

On searching the best mode for forex forecasting: bidirectional long short-term memory default mode is not enough

Hansun

Putri

Hugeng

2022

IJ-AI

View full text Add to dashboard Cite

show abstract

“…The sliding window technique involves partitioning a time series data sequence into two segments: the initial segment serves as the input window values, while the subsequent segment represents the predicted values [31]. This iterative process advances step by step, shifting one step at a time through the dataset, acquiring multiple samples from the training set.…”

Section: Sliding Window Methods (Time Steps)mentioning

confidence: 99%

Application of Recurrent Neural Network Bi-Long Short-Term Memory, Gated Recurrent Unit and Bi-Gated Recurrent Unit for Forecasting Rupiah Against Dollar (USD) Exchange Rate

Fayyad,

Kurniawan,

Anugrah

et al. 2024

PREDATECS

View full text Add to dashboard Cite

Foreign exchange rates have a crucial role in a country's economic development, influencing long-term investment decisions. This research aims to forecast the exchange rate of Rupiah to the United States Dollar (USD) by using deep learning models of Recurrent Neural Network (RNN) architecture, especially Bi-Long Short-Term Memory (Bi-LSTM), Gated Recurrent Unit (GRU), and Bi-Gated Recurrent Unit (Bi-GRU). Historical daily exchange rate data from January 1, 2013 to November 3, 2023, obtained from Yahoo Finance, was used as the dataset. The model training and evaluation process was performed based on various parameters such as optimizer, batch size, and time step. The best model was identified by minimizing the Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). Among the models tested, the GRU model with Nadam optimizer, batch size 16, and timestep 30 showed the best performance, with MSE 3741.6999, RMSE 61.1694, MAE 45.6246, and MAPE 0.3054%. The forecast results indicate a strengthening trend of the Rupiah exchange rate against the USD in the next 30 days, which has the potential to be taken into consideration in making investment decisions and shows promising economic growth prospects for Indonesia.

show abstract

“…Therefore, a proposed new CNN architecture incorporates a modular topology inspired by Tzilivaki et al [50], formulating a convolutional, orthogonal recurrent MCD replacing the pooling layers, followed by dense layers flattening their outputs. Compared with a typical CNN time series composed of convolutional, pooling, flattened, and dense layers, the proposed new CNN could enhance the robustness and forecasting performance of the Forex market [79].…”

Section: Proposed Novel Bio-inspired Model In Predicted Forex Market ...mentioning

confidence: 99%

Predicting Forex Currency Fluctuations Using a Novel Bio-Inspired Modular Neural Network

Bormpotsis,

Sedky,

Patel

2023

BDCC

View full text Add to dashboard Cite

In the realm of foreign exchange (Forex) market predictions, Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) have been commonly employed. However, these models often exhibit instability due to vulnerability to data perturbations attributed to their monolithic architecture. Hence, this study proposes a novel neuroscience-informed modular network that harnesses closing prices and sentiments from Yahoo Finance and Twitter APIs. Compared to monolithic methods, the objective is to advance the effectiveness of predicting price fluctuations in Euro to British Pound Sterling (EUR/GBP). The proposed model offers a unique methodology based on a reinvigorated modular CNN, replacing pooling layers with orthogonal kernel initialisation RNNs coupled with Monte Carlo Dropout (MCoRNNMCD). It integrates two pivotal modules: a convolutional simple RNN and a convolutional Gated Recurrent Unit (GRU). These modules incorporate orthogonal kernel initialisation and Monte Carlo Dropout techniques to mitigate overfitting, assessing each module’s uncertainty. The synthesis of these parallel feature extraction modules culminates in a three-layer Artificial Neural Network (ANN) decision-making module. Established on objective metrics like the Mean Square Error (MSE), rigorous evaluation underscores the proposed MCoRNNMCD–ANN’s exceptional performance. MCoRNNMCD–ANN surpasses single CNNs, LSTMs, GRUs, and the state-of-the-art hybrid BiCuDNNLSTM, CLSTM, CNN–LSTM, and LSTM–GRU in predicting hourly EUR/GBP closing price fluctuations.

show abstract

Comparative analysis of the application of Deep Learning techniques for Forex Rate prediction

Cited by 6 publications

References 8 publications

On searching the best mode for forex forecasting: bidirectional long short-term memory default mode is not enough

On searching the best mode for forex forecasting: bidirectional long short-term memory default mode is not enough

Application of Recurrent Neural Network Bi-Long Short-Term Memory, Gated Recurrent Unit and Bi-Gated Recurrent Unit for Forecasting Rupiah Against Dollar (USD) Exchange Rate

Predicting Forex Currency Fluctuations Using a Novel Bio-Inspired Modular Neural Network

Contact Info

Product

Resources

About