摘要Time series prediction methods based on conventional neural networks do not take into account the functional relations between the discrete observed values in the time series. This usually causes a low prediction accuracy. To solve this problem, a functional time series prediction model based on a process neural network is proposed in this paper. A Levenberg-Marquardt learning algorithm based on the expansion of the orthonormal basis functions is developed to train the proposed functional time series prediction model. The efficiency of the proposed functional time series prediction model and the corresponding learning algorithm is verified by the prediction of the monthly mean sunspot numbers. The comparative test results indicate that process neural network is a promising tool for functional time series prediction.
Abstract：Time series prediction methods based on conventional neural networks do not take into account the functional relations between the discrete observed values in the time series. This usually causes a low prediction accuracy. To solve this problem, a functional time series prediction model based on a process neural network is proposed in this paper. A Levenberg-Marquardt learning algorithm based on the expansion of the orthonormal basis functions is developed to train the proposed functional time series prediction model. The efficiency of the proposed functional time series prediction model and the corresponding learning algorithm is verified by the prediction of the monthly mean sunspot numbers. The comparative test results indicate that process neural network is a promising tool for functional time series prediction.
DING Gang;LIN Lin;ZHONG Shi-Sheng. Functional Time Series Prediction Using Process Neural Network[J]. 中国物理快报, 2009, 26(9): 90502-090502.
DING Gang, LIN Lin, ZHONG Shi-Sheng. Functional Time Series Prediction Using Process Neural Network. Chin. Phys. Lett., 2009, 26(9): 90502-090502.
[1] Podobnik B and Stanley H E 2008 Phys. Rev. Lett. 100 084102 [2] Scafetta N and West B J 2004 Phys. Rev. Lett. 92 138501 [3] Zhao L and Yang Y P 2009 Expert Syst. Appl. 36 2805 [4] Funahashi F 1989 Neural Networks 2 359 [5] Hornik K, Stinchcombe M and White H Neural Networks 2 183 [6] Eisenstein E, Kanter I, Kessler D A and Kinzel W 1995 Phys. Rev. Lett. 74 6 [7] Freking A, Kinzel W and Kanter I 2002 Phys. Rev. E 65 050903 [8] Kulkami D R, Parikh J C and Pratap R 1997 Phys.Rev. E 55 4508 [9] Ding G and Zhong S S 2006 Neural Network World16 15 [10] Ghazali R, Hussain A J, Liatsis P and Tawfik H 2008 Neural Comput. Appl. 17 311 [11] Glendinning R H and Fleet S L 2007 Signal Process. 87 79 [12] Antoniadis A, Paparoditis E and Sapatinas T 2009 Stat. Probab. Lett. 79 733 [13] Aneiros-Pereza G and Vieub P 2008 J. Multivar.Anal. 99 834 [14] Gatet L, Tap-Beteille H and Bony F 2009 IEEE Trans.Neu ral Networks 20 460 [15] Zhang L I, Tao H W, Holt C E, Harris W A and Poo M M 1998 Nature 395 37 [16] He X G and Liang J Z 2000 Chin. Engin. Sci. 240 (in Chinese) [17] Hagan M T and Menhaj M 1994 IEEE Trans. NeuralNetworks 5 989
2009, Vol. 26 
