Selection of Input Variables of Nonlinear Autoregressive Neural Network Model for Time Series Data Forecasting

Hermansah, Hermansah; Rosadi, Dedi; Abdurakhman, Abdurakhman; Utami, Herni

doi:10.14710/medstat.13.2.116-124

Cited by 6 publications

(6 citation statements)

References 12 publications

(12 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, if the condition > 3 is not met, the 1:4 lag is used. For descriptions and implementation details of determining input variables, see [5,6,8]. Furthermore, the number of neurons in the hidden layer is half of the number of input variables.…”

Section: Multi-output Narnn Modelmentioning

confidence: 99%

“…However, if the condition m > 3 is not met, 1:4 lag is used. The [5] was further extended in [6] by implementing a combination of learning algorithms, activation functions, and ensemble operators. The [6] approach was used the stepwise selection to obtain the most suitable NARNN model on univariate data.…”

Section: Introductionmentioning

confidence: 99%

“…The [5] was further extended in [6] by implementing a combination of learning algorithms, activation functions, and ensemble operators. The [6] approach was used the stepwise selection to obtain the most suitable NARNN model on univariate data. As an extension, this research was conducted, and the number of neurons in the hidden layer is half of the number of input variables.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multivariate Time Series Data Forecasting Using Multi-Output NARNN Model

Hermansah¹,

Rosadi²,

Abdurakhman³

et al. 2021

Advances in Social Science, Education and Humanities Research

Self Cite

View full text Add to dashboard Cite

This research proposes the multi-output Nonlinear Autoregressive Neural Network (NARNN) method to forecast multivariate time series data containing the input layer, one hidden layer, and the output layer. The multi-output NARNN method is performed by applying the logistic activation function and the resilient backpropagation learning algorithm. The stage of determining the input variable is chosen based on the number of data frequencies. The number of neurons in the hidden layer is half of the number of input variables. Simulation and empirical studies are conducted to test whether the proposed method works well for multivariate time series data forecasting. The simulation results show that the best performance is the simulation data generated from the MESTAR nonlinear model. The simulation study results are as expected. Empirical studies on Indonesia's inflation and Bank Indonesia interest rate data show that the multioutput NARNN method provides better forecasting accuracy than the VAR, VMA, and VARMA methods with a total MSE value of 0.054655 and a total MAPE of 0.026853 in the testing data.

show abstract

Section: Multi-output Narnn Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multivariate Time Series Data Forecasting Using Multi-Output NARNN Model

Hermansah¹,

Rosadi²,

Abdurakhman³

et al. 2021

Advances in Social Science, Education and Humanities Research

Self Cite

View full text Add to dashboard Cite

show abstract

“…This concept is classified into linear and nonlinear methods, with the popular univariate types being exponential smoothing and autoregressive integrated moving average (ARIMA) models. These methods have been reported to be successful in forecasting linear time-series data, and very poor based on designing nonlinear and complex parameters [1]. Meanwhile, the nonlinear forecasting generally provides irregular function specification requirements.…”

Section: Introductionmentioning

confidence: 99%

“…The studies in [8], [11]- [13] used the NAR model with sigmoid and linear activation functions at the hidden and output layers, with the backpropagation learning algorithm applied to update the parameters. Specifically, the non-automatic version of the forecasting algorithm based on the NAR model was observed in [1]. Forecasting time series with NAR is also found to be possible, by using multiple univariate models.…”

Section: Introductionmentioning

confidence: 99%

Automatic time series forecasting using nonlinear autoregressive neural network model with exogenous input

et al. 2021

Self Cite

View full text Add to dashboard Cite

This study aims to determine an automatic forecasting method of univariate time series, using the nonlinear autoregressive neural network model with exogenous input (NARX). In this automatic setting, users only need to supply the input of time series. Then, an automatic forecasting algorithm sets up the appropriate features, estimate the parameters in the model, and calculate forecasts, without the users’ intervention. The algorithm method used include preprocessing, tests for trends, and the application of first differences. The time series were tested for seasonality, and seasonal differences were obtained from a successful analysis. These series were also linearly scaled to [−1, +1]. The autoregressive lags and hidden neurons were further selected through the stepwise and optimization algorithms, respectively. The 20 NARX models were fitted with different random starting weights, and the forecasts were combined using the ensemble operator, in order to obtain the final product. This proposed method was applied to real data, and its performance was compared with several available automatic models in the literature. The forecasting accuracy was also measured by mean squared error (MSE) and mean absolute percent error (MAPE), and the results showed that the proposed method outperformed the other automatic models.

show abstract