Recurrent neural networks and time series prediction

Connor, Jerome T.; Atlas, Les

doi:10.1109/ijcnn.1991.155194

Cited by 65 publications

(27 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This allows RNNs to learn extensively complex sequential patterns. Studies by Connor and Atlas (1993), Logar et al (1993), Adam et al (1994), and Kamijo and Tanigawa (1990) show that performance of RNN is expected to be superior to that of ANN. However, the main drawback associated with RNNs is that they need more time to learn (estimate) than the standard neural networks because, in RNNs, output passes through the network more than once (depending on the type of the RNN) before the final output from the model is attained.…”

Section: Recurrent Neural Networkmentioning

confidence: 91%

Testing Forecast Accuracy of Foreign Exchange Rates: Predictions from Feed Forward and Various Recurrent Neural Network Architectures

Kiani

Kastens

2008

Comput Econ

View full text Add to dashboard Cite

show abstract

Section: Recurrent Neural Networkmentioning

confidence: 91%

Testing Forecast Accuracy of Foreign Exchange Rates: Predictions from Feed Forward and Various Recurrent Neural Network Architectures

Kiani

Kastens

2008

Comput Econ

View full text Add to dashboard Cite

show abstract

“…Their superiority against feedfoward networks when performing nonlinear time series prediction is well documented in Connor et al (1993) and Adam et al (1994). In financial applications, Kamijo et al (1990) applied them successfully to the recognition of stock patterns of the Tokyo stock exchange while Tenti (1996) achieved remarkable results using RNNs to forecast the exchange rate of the Deutsche Mark.…”

Section: Literature Reviewmentioning

confidence: 93%

GP algorithm versus hybrid and mixed neural networks

Dunis

Laws

Karathanasopoulos

2013

The European Journal of Finance

View full text Add to dashboard Cite

In the current paper we present an integrated genetic programming environment, called java GP Modelling. The java GP Modelling environment is an implementation of the steady-state genetic programming algorithm. That algorithm evolves tree based structures that represent models of inputoutput relation of a system. The motivation of this paper is to compare the GP algorithm with neural network architectures when applied to the task of forecasting and trading the ASE 20 Greek Index using only autoregressive terms as inputs. This is done by benchmarking the forecasting performance of the GP algorithm and 6 different ARMA-Neural Network combination designs representing a Hybrid, Mixed Higher Order Neural Network (HONN), a Hybrid, Mixed Recurrent Network (RNN), a Hybrid, Mixed classic Multilayer Perceptron (MLP) with some traditional techniques, either statistical such as a an autoregressive moving average model (ARMA), or technical such as a moving average convergence/divergence model (MACD), plus a naïve trading strategy. More specifically, the trading performance of all models is investigated in a forecast and trading simulation on ASE 20 time series closing prices over the period 2001-2008 using the last one and a half years for out-of-sample testing. We use the ASE 20 daily series as many financial institutions are ready to trade at this level and it is therefore possible to leave orders with a bank for business to be transacted on that basis.As it turns out, the GP model does remarkably well and outperforms all other models in a simple trading simulation exercise. This is also the case when more sophisticated trading strategies using confirmation filters and leverage are applied, as the GP model still produces better results and outperforms all other neural network and traditional statistical models in terms of annualised return.

show abstract

“…Another neural model is the recurrent neural network (RNN) which can learn sequences as time evolves and responds to the same input pattern differently at different time, depending on the previous input patterns as well. Both neural network models represent the performance for nonlinear prediction of time series (Connor, Martin and Atlas, 1994;Connor and Atlas, 1991). Sometimes recurrent neural network (RNN) lead to improved forecasting performance as compared to MFNN (Ho, Xie and Goh, 2002;Connor and Atlas, 1991).…”

Section: Introductionmentioning

confidence: 99%

Comparing forecasting performances between multilayer feedforward neural network and recurrent neural network in Malaysia’s load

Mohamed¹,

Ahmad²,

Ismail³

et al. 2010

Journal of Interdisciplinary Mathematics

View full text Add to dashboard Cite

This paper presents the use of two artificial neural networks models, namely the multilayer feedforward neural network (MLFF) and the recurrent neural network (RNN) are applied for Malaysia's load forecasting. For this purpose, a half hourly load data is divided equally into three distinct sets for training, validation and testing. We use backpropagation as the learning algorithm and the sigmoid function as the transfer function for both hidden land output layers. The forecasting performances of were compared between these two models. We use the sum squared error (SSE) as the measure of performance and the correlation coefficient r , as the measure of relationship between the actual and the predicted values. Results show that, multilayer feedforward neural network (MLFF) and recurrent neural network (RNN) have comparable accuracy but the sum squared error for multilayer feedforward neural network (MLFF) is lower, thus making it better model than recurrent neural network (RNN).

show abstract

Recurrent neural networks and time series prediction

Cited by 65 publications

References 4 publications

Testing Forecast Accuracy of Foreign Exchange Rates: Predictions from Feed Forward and Various Recurrent Neural Network Architectures

Testing Forecast Accuracy of Foreign Exchange Rates: Predictions from Feed Forward and Various Recurrent Neural Network Architectures

GP algorithm versus hybrid and mixed neural networks

Comparing forecasting performances between multilayer feedforward neural network and recurrent neural network in Malaysia’s load

Contact Info

Product

Resources

About