Nonlinear parametric models for financial time series

The thesis is dedicated to study of nonlinear parametric models for financial time series. It contains the summary of basic terms of this issues in brief. The next part is dedicated to the survey of different linear and nonlinear models with description of their basic features. Threshold autoregressive model and bilinear model are presented in details. For these two models, basic features, tests for linearity and estimations are introduced. The practical part is based on the theory described in previous chapters. Particular tests for linearity for both models and estimated instruments are analysed on simulated and real data