Threshold models in non-linear time series analysis. Lecture notes in statistics, No.21

One Introduction.- 1. Time Series Model Building.- 2. Stationarity.- 3. Linear Gaussian Models.- 4. Some Advantages and Some Limitations of Arma Models.- 5. What Next?.- Two Some Basic Concepts.- 1. Orientation.- 2. Limit Cycles.- 3. Some Examples of Threshold Models.- 4. Time Delay.- 5. Discussion.- Three Threshold Models.- 1. A Canonical Form.- 2. Generality of Setar Models.- 3. Non-Linear Difference Equations.- 4. Threshold Models and Discrete-Time Non-Linear Vibrations.- 5. Ergodicity.- 6. Stationary Distributions and Moments.- 7. Cyclical Structure and Multi-Step-Ahead Forecasting.- Four Identification.- 1. A General Principle.- 2. Estimation of Parameters.- 3. Sampling Properties.- 3.1 A General Result.- 3.2 Application of General Result to SETAR.- 3.3 Some Simulation Studies.- 3.4 Standard Errors of Some Parameter Estimates.- 3.5 Unsolved Problems.- 4. Diagnostics and Graphical Methods.- 5. Miscellanea.- Five Some Case Studies.- 1. Analysis of Some Ecological Data.- 1.1 Ecological Background.- 1.2 Some Graphical Results.- 1.3 A Full Identification.- 1.4 Diagnostics.- 1.5 Transformation.- 1.6 Extension to Two Species.- 2. Analysis of the Sunspot Numbers.- 2.1 Some Background.- 2.2 SETAR Modelling.- 2.3 Transformation.- 2.4 Multi-step-ahead Forecasting.- 2.5 Some Discussion.- 3. Analysis of Some Riverflow Data.- 3.1 Some Background.- 3.2 TARSO Modelling of Vatnsdalsa River, Iceland (1972-1974).- 3.3 TARSO Modelling of Jokulsa River, Iceland (1972-1974).- 3.4 Some Discussion.- 4. A Case Study with Laboratory Data.- 5. A Fuzzy Extension.- 6. Concluding Remarks.- Appedices.- References.- Author Index.