Feature selection for time series prediction - A combined filter and wrapper approach for neural networks

NEUROCOMPUTING

卷:73

Article

Crone, Sven F.¹  Kourentzes, Nikolaos¹ 

[1] Univ Lancaster, Sch Management, Dept Management Sci, Ctr Forecasting, Lancaster LA1 4YX, England

关键词: Time series prediction;    Forecasting;    Artificial neural networks;    Automatic model specification;    Feature selection;    Input variable selection;   

DOI  :  10.1016/j.neucom.2010.01.017

来源: Elsevier

【 摘 要 】

Modelling artificial neural networks for accurate time series prediction poses multiple challenges, in particular specifying the network architecture in accordance with the underlying structure of the time series. The data generating processes may exhibit a variety of stochastic or deterministic time series patterns of single or multiple seasonality, trends and cycles, overlaid with pulses, level shifts and structural breaks, all depending on the discrete time frequency in which it is observed. For heterogeneous datasets of time series, such as the 2008 ESTSP competition, a universal methodology is required for automatic network specification across varying data patterns and time frequencies. We propose a fully data driven forecasting methodology that combines filter and wrapper approaches for feature selection, including automatic feature evaluation, construction and transformation. The methodology identifies time series patterns, creates and transforms explanatory variables and specifies multilayer perceptrons for heterogeneous sets of time series without expert intervention. Examples of the valid and reliable performance in comparison to established benchmark methods are shown for a set of synthetic time series and for the ESTSP'08 competition dataset, where the proposed methodology obtained second place. (C) 2010 Elsevier B.V. All rights reserved.

【 授权许可】

Free