I need some Ideas. How fitting is non-linearity method for analyzing financial data? What is the advantage of this method in analyzing the financial data? do you all have any idea? kindly share....it will be helpful if you could suggest any paper on this issue...thank you and have a nice day.
It is very appropiate because it allows for the existence of different states of the world and for the dynamics to be different. You can extend AR models to allow for nonlinear behavior, the resulting models are easy to understand and interpret.
You can also find useful information in this book, Granger and Ter¨asvirta
(1993).
One particular reference by Granger (1993) states that share prices are non-linear. Before attempting to fit a non-linear model, there needs to be both an intuitive reason and evidence that a linear model will not work. Having established that a linear model is not suitable for financial variables, your next question becomes one of finding the correct model fit. The best bet here would be time series models. If there is no intuition on model fit, then a last resort could be to use neural networks. These are non-linear models that do not have a specific form. They are quite accurate at prediction, but lack interpretation.
Basically I would say it depends on the type of data and relationships you examine. I guess that the best way to 'check' if any there are non-linearities emerging, just plot the data and see. You can also use specific statistical test (Wald test). best.