Nonlinearity Tests For Bilinear Time Series Data

Kertas kerja ini membincangkan dua ujian tak linear bagi data siri masa iaitu Ujian Keenan dan Ujian–F. Ujian–ujian ini adalah berdasarkan pendekatan domain masa dan menggunakan pengiraan yang tidak kompleks jika dibandingkan dengan pendekatan domain frekuensi. Kedua–dua ujian sangat sesuai digunakan ke atas data bilinear disebabkan keduanya boleh diungkapkan ke bentuk kembangan Volterra. Di dalam kajian ini, kami membangunkan program khas untuk ujian–ujian di dalam pakej S–Plus 2000. Kami akan menunjukkan melalui kajian simulasi bahawa ujian–ujian tersebut berfungsi dengan baik untuk membezakan data linear dan data tak linear daripada model siri masa bilinear. Ujian–ujian tak linear tersebut telah digunakan ke atas empat set data yang mengandungi ciri–ciri tak linear dan keputusan yang dihasilkan adalah baik. Kata kunci: Ujian Keenan, Ujian-F, ujian tak linear, bilinear This paper discusses two nonlinearity tests in time series analysis, which are the Keenan’s test and F–test. The tests are based on time domain approach and are computationally less complex than the frequency domain approach. Both tests are especially suitable for data generated from bilinear model as both can be expressed in Volterra expansion form. In this study, programs for both tests are developed in S–Plus 2000 package. Through simulation studies, it will be shown that the tests work well to distinguish linear from nonlinear data set generated from bilinear model. The nonbilinearity tests were applied on four real data sets which have nonlinear characteristics and the results obtained are desirable. Key words: Keenan’s test, F-test, nonlinearity test, bilinea

[1]  J. Tukey.  One Degree of Freedom for Non-Additivity , 1949 .

[2]  J. Yuan.  Testing Linearity For Stationary Time Series Using the Sample Interquartile Range , 2000 .

[3]  M. Hinich.  Testing for Gaussianity and Linearity of a Stationary Time Series. , 1982 .

[4]  R Furth,et al.  Non-Linear Problems in Random Theory , 1960 .

[5]  H. Tong,et al.  On tests for non‐linearity in time series analysis , 1986 .

[6]  Neville Davies,et al.  A portmanteau test for self-exciting threshold autoregressive-type nonlinearity in time series , 1986 .

[7]  Cathy W. S. Chen.  Bayesian inferences and forecasting in bilinear time series models , 1992 .

[8]  C. Granger,et al.  An introduction to bilinear time series models , 1979 .

[9]  George E. P. Box,et al.  Time Series Analysis: Forecasting and Control , 1977 .

[10]  R. Fildes.  Journal of the Royal Statistical Society (B): Gary K. Grunwald, Adrian E. Raftery and Peter Guttorp, 1993, “Time series of continuous proportions”, 55, 103–116.☆ , 1993 .

[11]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1971 .

[12]  R. Tsay.  Nonlinearity tests for time series , 1986 .

[13]  D. M. Keenan,et al.  A Tukey nonadditivity-type test for time series nonlinearity , 1985 .

[14]  T. Rao,et al.  An Introduction to Bispectral Analysis and Bilinear Time Series Models , 1984 .

[15]  H. Tong,et al.  Threshold Autoregression, Limit Cycles and Cyclical Data , 1980 .

[16]  Berlin Wu,et al.  A fuzzy identification procedure for nonlinear time series: With example on ARCH and bilinear models , 1999, Fuzzy Sets Syst..

[17]  Howell Tong,et al.  Threshold autoregression, limit cycles and cyclical data- with discussion , 1980 .

[18]  T. Rao,et al.  A TEST FOR LINEARITY OF STATIONARY TIME SERIES , 1980 .