A BP Neural Network Activation Function Used in Exchange Rate Forecasting

Abstract Deep learning has substantially advanced the state of the art in computer vision, natural language processing, and other fields. The paper examines the potential of deep learning for exchange rate forecasting. We systematically compare long short-term memory networks and gated recurrent units to traditional recurrent network architectures as well as feedforward networks in terms of their directional forecasting accuracy and the profitability of trading model predictions. Empirical results indicate the suitability of deep networks for exchange rate forecasting in general but also evidence the difficulty of implementing and tuning corresponding architectures. Especially with regard to trading profit, a simpler neural network may perform as well as if not better than a more complex deep neural network.

Download Full-text

Exchange Rate Forecasting Method Based on Particle Swarm Optimization and Probabilistic Neural Network Model

2011 International Conference on Network Computing and Information Security ◽

10.1109/ncis.2011.65 ◽

2011 ◽

Author(s):

Bingxiang Liu ◽

Hua Wang ◽

Xiang Cheng

Keyword(s):

Neural Network ◽

Particle Swarm Optimization ◽

Exchange Rate ◽

Network Model ◽

Neural Network Model ◽

Probabilistic Neural Network ◽

Particle Swarm ◽

Exchange Rate Forecasting ◽

Swarm Optimization ◽

Forecasting Method

Download Full-text

PEMODELAN WAVELET NEURAL NETWORK UNTUK PREDIKSI NILAI TUKAR RUPIAH TERHADAP DOLAR AS

Jurnal Gaussian ◽

10.14710/j.gauss.v9i2.27823 ◽

2020 ◽

Vol 9 (2) ◽

pp. 217-226

Author(s):

Tri Yani Elisabeth Nababan ◽

Budi Warsito ◽

Agus Rusgiyono

Keyword(s):

Neural Network ◽

Exchange Rate ◽

Wavelet Transforms ◽

Activation Function ◽

Wavelet Neural Network ◽

Training Data ◽

The Difference ◽

Combination Of Methods ◽

Hidden Layer ◽

The U.S

Each country has its own currency that is used as a tool of exchange rate valid in the transaction process. In the process of transaction between countries often experience problems in terms of payment because of the difference in the value of money prevailing in each country. The price movement of the exchange rate or the value of foreign currencies that fluctuate from time to time it encouraged predictions of the value of the rupiah exchange rate against the U.S. dollar. Wavelet Neural Network (WNN) is a combination of methods between wavelet transforms and Neural networks. WNN modeling begins with wavelet decomposition resulting in wavelet coefficients and scale coefficients. Selection of inputs is based on PACF plots and divides into training data and testing data. To determine the final output by calculating the value of MAPE in data testing. The best architecture on WNN model for prediction of the value of the rupiah exchange rate against the U.S. dollar is a model with sigmoid logistic activation function, 2 neurons in the input layer, 10 neurons in the hidden layer, and 1 neuron in the output layer. The MAPE value is obtained at 0.2221%.

Download Full-text