Strategies for multi-step-ahead available parking spaces forecasting based on wavelet transform

Abstract

A new methodology for multi-step-ahead forecasting was proposed herein which combined the wavelet transform (WT), artificial neural network (ANN) and forecasting strategies based on the changing characteristics of available parking spaces (APS). First, several APS time series were decomposed and reconstituted by the wavelet transform. Then, using an artificial neural network, the following five strategies for multi-step-ahead time series forecasting were used to forecast the reconstructed time series: recursive strategy, direct strategy, multi-input multi-output (MIMO) strategy, DIRMO strategy (a combination of the direct and MIMO strategies), and newly proposed recursive multi-input multi-output (RECMO) strategy which is a combination of the recursive and MIMO strategies. Finally, integrating the predicted results with the reconstructed time series produced the final forecasted available parking spaces. Three findings appear to be consistently supported by the experimental results. First, applying the wavelet transform to multi-step ahead available parking spaces forecasting can effectively improve the forecasting accuracy. Second, the forecasting resulted from the DIRMO and RECMO strategies is more accurate than that of the other strategies. Finally, the RECMO strategy requires less model training time than the DIRMO strategy and consumes the least amount of training time among five forecasting strategies.

Keywords

available parking spaces / multi-step ahead time series forecasting / wavelet transform / forecasting strategies / recursive multi-input multi-output strategy

References

[1] YangZ-s, LiuH-h, WangX-yue. The research on the key technologies for improving efficiency of parking guidance system [J]. Intelligent Transportation Systems, 2003, 2(2): 1177-1182

[2] LiuS-x, GuanH-z, YanH, YinH-huan. Unoccupied parking space prediction of chaotic time series [C]//. Tenth International Conference of Chinese Transportation Professionals, 201021222131

[3] JiY-j, TangD-n, GuoW-h, BlytheP T, WangWei. Forecasting available parking space with largest Lyapunov exponents method [J]. Journal of Central South University, 2014, 21(4): 1624-1632

[4] TaiebS B, BontempiG, AtiyaA F, SorjamaaA. A review and comparison of strategies for multi-step ahead time series forecasting based on the NN5 forecasting competition [J]. Expert Systems with Applications, 2012, 39(8): 7067-7083

[5] GuoZ-h, ZhaoW-g, LuH-y, WangJ-zhou. Multi-step forecasting for wind speed using a modified EMD-based artificial neural network model [J]. Renewable Energy, 2012, 37(1): 241-249

[6] AmendolaA, NiglioM, VitaleC. Multi-step SETARMA predictors in the analysis of hydrological time series [J]. Physics and Chemistry of the Earth, Parts A/B/C, 2006, 31(18): 1118-1126

[7] ZhangX-r, HutchinsonJ. Simple architectures on fast machines: practical issues in nonlinear time series prediction [C]//. Time Series Prediction: Forecasting the Future and Understanding the Past, 1993, New Mexico, Santa Fe Institute: 2231

[8] SorjamaaA, HaoJ, ReyhaniN, JiY-n, LendasseA. Methodology for long-term prediction of time series [J]. Neurocomputing, 2007, 70(16): 2861-9

[9] Hamza EbiC, AkayD, KutayF. Comparison of direct and iterative artificial neural network forecast approaches in multiperiodic time series forecasting [J]. Expert Systems with Applications, 2009, 36(2): 3839-3844

[10] WeigendA S, HubermanB A, RumelhartD E. Predicting sunspots and exchange rates with connectionist networks [C]//. Nonlinear Modeling & Forecasting, 1991201237

[11] BirattariM, BontempiG, BersiniH. Lazy learning meets the recursive least squares algorithm [J]. Advances in neural Information Processing Systems, 1999375381

[12] KlineD M, ZhangG. Methods for multi-step time series forecasting with neural networks [J]. Neural Networks in Business Forecasting, 2004226250

[13] ChengH-b, TanP-n, GaoJ, ScrippsJ. Multistep-ahead time series prediction [C]//. Advances in Knowledge Discovery and Data Mining, 2006765774

[14] BontempiG. Long term time series prediction with multi-input multi-output local learning [C]//. Proc 2nd ESTSP, 2008145154

[15] BontempiG, TaiebS B. Conditionally dependent strategies for multiple-step-ahead prediction in local learning [J]. International journal of forecasting, 2011, 27(3): 689-699

[16] TaiebS B, BontempiG, SorjamaaA, LendasseA. Long-term prediction of time series by combining direct and mimo strategies [C]//. International Joint Conference on Neural Networks, 200930543061

[17] TaiebS B, SorjamaaA, BontempiG. Multiple-output modeling for multi-step-ahead time series forecasting [J]. Neurocomputing, 2010, 73(10): 1950-1957

[18] YangFei. Traffic flow predition model based on echo state networks and related research [D]. Beijing University of Posts and Telecommunivations, 2012

[19] IchirM M, Mohammad-DjafariA. Hidden Markov models for wavelet-based blind source separation [J]. Image Processing, IEEE Transactions on, 2006, 15(7): 1887-1899

[20] PaivaH M, GalvO R K H. Optimized orthonormal wavelet filters with improved frequency separation [J]. Digital Signal Processing, 2012, 22(4): 622-627

[21] MallatS G. A theory for multiresolution signal decomposition: the wavelet representation [J]. Pattern Analysis and Machine Intelligence, 1989, 11(7): 674-693

[22] TiaoG C, TsayR S. Some advances in non-linear and adaptive modelling in time-series [J]. Journal of Forecasting, 1994, 13(2): 109-131

[23] WeigendA S, GershenfeldN ATime series prediction: Forecasting the future and understanding the past [M], 1994, New Jersey, Addison Wesley

[24] KarunasingheD S, LiongS Y. Chaotic time series prediction with a global model: Artificial neural network [J]. Journal of Hydrology, 2006, 323(1): 92-105

[25] WuC-you. The research and application on neural network [D]. Northeast Agricultural University, 2007