This study proposes two metrics using the nearest neighbors method to improve the accuracy of
time-series forecasting. These two metrics can be treated as a hybrid forecasting approach to combine linear
and non-linear forecasting techniques. One metric redefines the distance in k-nearest neighbors based on the
coefficients of autoregression (AR) in time series. Meanwhile, an improvement to Kulesh’s adaptive metrics in
the nearest neighbors is also presented. To evaluate the performance of the two proposed metrics, three types
of time-series data, namely deterministic synthetic data, chaotic time-series data and real time-series data, are
predicted. Experimental results show the superiority of the proposed AR-enhanced k-nearest neighbors methods
to the traditional k-nearest neighbors metric and Kulesh’s adaptive metrics.
PAN Feng1* (潘 峰), ZHAO Hai-bo2 (赵海波), LIU Hua-shan1 (刘华山)
. Time-Series Forecasting Using Autoregression Enhanced k-Nearest Neighbors Method[J]. Journal of Shanghai Jiaotong University(Science), 2013
, 18(4)
: 434
-442
.
DOI: 10.1007/s12204-013-1418-x
[1] Makridakis S, Wheelwright S, Hyndman R. Forecasting: Methods and applications [M]. New York: Wiley, 1998.
[2] Kuo R J, Xue K C. Fuzzy neural networks with application to sales forecasting [J]. Fuzzy Sets and Systems, 1999, 108(2): 123-143.
[3] Wagner N, Michalewicz Z, Khouja M, et al. Time series forecasting for dynamic environments: The DyFor genetic program model [J]. IEEE Transactions on Evolutionary Computation, 2007, 11(4): 433-452.
[4] Kulesh M, Holschneider M, Kurennaya K. Adaptive metrics in the nearest neighbours method [J]. Physica D, 2008, 237(5): 283-291.
[5] Small M, Tse C K. Optimal embedding parameters: A modeling paradigm [J]. Physica D, 2004, 194(3-4): 283-296.
[6] Cao L J, Tay F E H. Support vector machine with adaptive parameters in financial time series forecasting [J]. IEEE Transaction on Neural Networks, 2003, 14(6): 1506-1518.
[7] Huang K, Yu T H K. Ratio-based lengths of intervals to improve fuzzy time series forecasting [J]. IEEE Transactions on Systems, Man, and Cybernetics. Part B: Cybernetics, 2006, 36(2): 328-340.
[8] Yu H K. A refined fuzzy time-series model for forecasting [J]. Physica A, 2005, 346(3-4): 657-681.
[9] Huang K, Yu H K. An N-th order heuristic fuzzy time series model for TAIEX forecasting [J]. International Journal of Fuzzy Systems, 2003, 5(4): 247-253.
[10] Sfetsos A, Siriopoulos C. Time series forecasting with a hybrid clustering scheme and pattern recognition [J]. IEEE Transactions on System, Man, and Cybernetics. Part A: Systems and Humans, 2004, 34(3): 399-405.
[11] Zhang G P. Time series forecasting using a hybrid ARIMA and neural network model [J]. Neurocomputing, 2003, 50(1): 159-175.
[12] Murray D B. Forecasting a chaotic time series using an improved metric for embedding space [J]. Physica D, 1993, 68(8): 318-325.
[13] Cao L. Practical method for determining the minimum embedding dimension of a scalar time series [J]. Physica D, 1997, 110(1-2): 43-50.
[14] Weigend A S, Gershenfeld N A. Results of the time series prediction competition at the Santa Fe Institute [C]// IEEE International Conference on Neural Networks. San Francisco, USA: IEEE, 1993: 1786-1793.
