Time series forecasting using neural networks: should the data be deseasonalized first?

By M. Nelson, T. Hill, W. Remus, Marcus O'Connor. 1999

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

Forecasting Across Time Series Databases using Long Short-Term Memory Networks on Groups of Similar Series (2017, methodology): “In particular, Nelson et al. (1999) compare the forecasts generated from NNs trained with deseasonalized data and non-deseasonalized data, using 68 monthly time series from the M-competition (Makridakis et al., 1982).”
Forecasting across time series databases using recurrent neural networks on groups of similar series: A clustering approach (2017, methodology): “Recent developments have been mainly around preprocessing techniques such as deseasonalization and detrending to supplement the NN’s learning process, and novel NN architectures such as recurrent neural networks (RNN), echo state networks (ESN), generalized regression neural networks (GRNN) and ense”
A comparative study of linear and nonlinear models for aggregate retail sales forecasting (2003, background): “Based on a sample of 68 time series from the same database, Nelson et al. (1999), however, concluded just the opposite.”
A review and comparison of strategies for multi-step ahead time series forecasting based on the NN5 forecasting competition (2011, unknown): “…Curry and Morgan, 2006; Anders and Korn, 1999; Chapelle and Vapnik, 2000), eﬀect of deseasonalization (Hylleberg, 1992; Makridakis et al., 1998; Nelson et al., 1999; Zhang and Qi, 2005), forecasts combination (Bates, J. M. and Granger, C. W. J., 1969; Clemen, 1989; Timmermann, 2006) and many…”
The use of data mining and neural networks for forecasting stock market returns (2005, methodology): “In general, large-scale deterministic components, such as trends and seasonal variations, should be eliminated from the inputs since the network will attempt to learn the trend and use it in the prediction (Nelson et al., 1999; Pantazopoulos et al., 1998).”