* Saarman@yahoocom Ali_r367@yahoocom 0075 07 MGN G7 C5 C58 C53 JEL
3 7 TAR 6 ARCH Chaos Theory Deterministic 3 Exponential Smoothing Autoregressive Integrated Moving Average 5 Autoregressive Conditional Heteroskedasticity 6 Bilinear 7 Threshold Auto Regressive
00 998 995 3 996 5 6 Zhang and Hu Artificial Neural Network 3 Prtugal Unobserved Component Models 5 Church and Curram 6 Dimension Correlation
003 98 007 3 5 8 7 007 6 0 9 008 Generalized Auto Regresive Conditional Heteroskedasticity Adaptive Neuro Fuzzy Inference System 3 Tully and lucey Threshold 5 Exponetial 6 Dunis and Nattani 7 Quantitive 8 Nearest Neighbors 9 Correct Directional Change 0 Risk Adjusted Return Parisi and Diaz Rooling Neural Network
009 0 38 35373 37380 387 3 33885 39 058 007 Fahimifard Yayar 3 Levenberg Marquardt Algorithm Takagi Sugeno
3 q d p p p, q p, q q 3 p, q = + + + + ~ 0, 008 q p p, d, q 3 98 ARCH ARCH 3 986 Volatility Clustering Engel 3 Bollerslev
ARCH E ARCHM p, q ARCHp I 986 ARCHq = + + q 986 ARCH 000 986 p q p, q 00 = + + + 6 3 33 ARCH in Mean Model Exponential 3 Integrated Johnston and Scott 5 Bouchad and Matacz 6 Soft Computing
TAR ARCH 00 00 =,, i t f = + + 5 w i b i0 b 0 w ij Exponential Smoothing Input Pattern
00 3 987 IID DBS, =,,, 6 m m,t W m,t W Scheinnkman, Dechert & Brock
=H 0 =H 0 =H 0 =H =H =H 39 Z %0 6 %5 Z 6
3 7 3 07 0075 366 95 3 World Gold Council wwwgoldorg Training Set 3 Test Set Out of sample
366 593 0765 976 586 56 00867 00553 89 0 5 Skewness Kurtosis
5 p, d, q 986 t 950 0000 3 MA AR 0 9 AR 0, 9 5 Granger and Newbold Augmented DickeyFuller test
3 0 3 5 6 7 8 9 0 0 3375 338 338 3386 339 3396 3396 3399 3398 3393 3386 338 338 338 339 338 3388 339 3396 3399 338 3386 338 3387 398 3306 33 3379 338 3386 336 3387 3388 3 3387 339 3305 3357 33 339 3387 0 3358 335 3356 339 3385 33 3369 3306 33 333 335 3369 3368 336 5 3396 3387 339 339 33 367 37 377 3379 337 335 6 3397 339 338 3387 333 37 375 38 369 330 3309 7 30 3398 3387 3389 33 376 358 339 3 330 378 8 303 3 3367 3367 3368 38 338 3 36 33 38 9 3398 3388 333 3358 337 38 3307 330 33 36 37 0 3398 30 3385 3363 3363 3368 379 33 33 396 37 AR MA 5
5, 6, 6 35
5 0 M= M=3 M= M=5 M=6 05 30 37 3 35 0 0 6 7 37 6 56 83 353 35 76 9 30 37 35 3 55 56 67 0 3 0 37 38 56 7 35 56 07 6 3 6 9 93 300 33 5
00 0 80 7 7 Matlab Sugeno fuzzy inference
6 99 MAE 5 TIC RMSE MAPE = 7 = 00 8 = 9 = + 0 993 7 008 6 Hykin Root Mean Squared Error RMSE 3 Mean Absolute Error MAE Mean Absolute Percentage Error MAPE 5 Theil inequality coefficient TIC 6 Brooks 7 Gerlow, Irowin and Liu
99 995 99 3 = Z t+s = If y t+s *f t,s >0 Z t+s =0 Otherwise T T t s f t,s t Y t = Z t+s = If y t+s y t f t,s y t >0 Z t+s =0 Otherwise 3 MGN Leitch and Tanner Pesaran and Timmerman 3 Correct Sign Predictions Correct Direction Change Predictions
00 e,t e,t D t S t =, +, =,,, =, = 5 x p x p x m MGN = 6 MGN N D t S t sd 00 N t 5 CSP CDCP TIC MAPE MAE RMSE 08 003 08 337 0 688 060 0, 9 090 050 0708 93 060 036 060 065 06 68 0085 033 Diebold and Mariano
65 8 6 5 0 MGN 6 39503 03 0935 ARAM 7
CDCP CSP TIC MAPE MAE RMSE MGN
ANN 38 975 39 39 387 5 Bollerslev, T 986 Generalized Autoregressive Conditional Heteroskedasticity Journal of Economics; 3 30737 6 Brock, W A, W Dechert, J Scheinkman 987 A Test for Independence Based on the Correlation Dimension Working paper, University of Winconsin at Madison, University of Houston, and University of Chicago 7 Brooks, C 008 Introductory Econometrics for Finance Second Edition, Cambridge University Press 8 Church, k b, Curram S P 996 Forecasting consumers' expenditure A comparison between econometric and neural network models International Journal of Forecasting,,, 55 67 9 Diebold F, Mariano R 00 Comparing predictive accuracy Journal of Business and Economic Statistics 0, 3 0 Engle F R 98 Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation Econometrica, 987007 Fahimifard, S M, Salarpour, M, Sabouhi, M, Shirzady, S 009 Application of to Agricultural Economic Variables Forecasting Case Study Poultry Retail Price, Artificial Intelligence, 657
Gerlow, M, Irwin, S, Liu, T 993 Economic Evaluation of Commodity Price Forecasting Models International Journal of Forecasting, 9387 3 Granger, C W J, Newbold, P 986 Forecasting economic time series Orlando Academic Press Gujarati D 008 Basic Econometrics 5 th Edition McGrawHill 5 Haykin, S 99, Neural Networks a Comprehensive Foundation Macmillan, New York 6 httpwwwgoldorg 7 Johnston k, Scott, E 000 Models and The Stochastic Process Underlying Exchange Rate Price Change Journal of Financial and Strategic Decisions, Vol pp 3 8 Leitch, G and Tanner, JE 99 Economic Forecast Evaluation Profits versus the Conventional Error Measures The American Economic Review AER, 83, 580 90 9 Parisi, A, Parisi, F, Díaz, D 008 Forecasting gold price changes Rolling and recursive neural network models, Journal of Multinational Financial Management, Elsevier, vol 85, pages 7787, December 0 Pesaran, M H, Timmerman, A 99 A Simple Nonparametric Test of Predictive Performance journal of Business & Economic Statistics 0,, 665 Prtugal, N S 995, Neural networks versus time series Methods A Forecasting Exercises, th International Symposium on Forecasting, Sweden Tully, E, Lucey, B M 007 A power examination of the gold market Research in international business and finance, 3630 3 Yayar, M, Hekim, M, Yelmaz, V, Bakirci, F 0 A comparison of and techniques in the forecasting of electric energy consumption of Tokat province in Turkey Zarranezhad, M, Raoofi, A, kiyani, p 0 Evaluation and comparison of performance of and in forecasting the daily gold prices, The First international conference on econometrics and methods applications 5 Zhang, G, Hu, MY 998 Neural Network Forecasting of the British PoundUS Dollar Exchange Rate, International Journal of Management Science, 6 95 506 6 Zhang, G, Hu, MY 00 A Simulation Study of Artificial Neural Networks for Nonliner Time Series Forecasting Comput Oper Res 8, pp 38 396 7 Zhang, G, Hu, MY 00 Forecasting with Artificial Neural Networks The state of the Art, Neurocomputing 56,pp 05 3