Neural modeling for time series: A statistical stepwise method for weight elimination

By M. Cottrell, B. Girard, Y. Girard, M. Mangeas, C. Muller. 1995

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Support vector machines experts for time series forecasting (2003, methodology): “Sunspot data The sunspot data set has long served as a benchmark and been well studied in the previous literature [5] [7] [8] [17] [18] [27] [30] [32].”
A comparison of univariate methods for forecasting electricity demand up to a day ahead (2006, background): “Neural networks will perform poorly if the time series is nonstationary ( Cottrell, Girard, Girard, Mangeas, & Muller, 1995 ).”
A simulation study of artificial neural networks for nonlinear time-series forecasting (2001, unknown): “[33], and Schitienkopf et al.”
A dynamic artificial neural network model for forecasting time series events (2005, methodology): “Cottrell et al. (1995) report the results of applying regression and traditional neural network models to the sunspot data.”
Time series forecasting using a hybrid ARIMA and neural network model (2003, unknown): “[7].”
Avoiding Pitfalls in Neural Network Research (2007, unknown): “For example, autoregressive models can be implemented via neural networks [31].”
Neural networks for short-term load forecasting: a review and evaluation (2001, methodology): “Reference [38] studied the accuracy of multi-steps forecasts obtained iteratively by a MLP on the -competition time series, and found that this MLP outperformed the statistical models; however, [ 19 ], [20] reported that the MLP outputs may behave chaotically.”
Neural networks in geophysical applications (2000, methodology): “Other pruning techniques can be found in Karnin (1990), Pellilo and Fanelli (1993), and Cottrell et al. (1995).”
An artificial neural network (p, d, q) model for timeseries forecasting (2010, methodology): “The neural network model used is composed of four inputs, four hidden and one output neurons ðN 4 1Þ ), as also employed by Cottrell et al. (1995), De Groot and Wurtz (1991) and Zhang (2003). Two forecast horizons of 35 and 67 periods are used in order to assess the forecasting performance of models”