Validation of volatility models
Article Abstract:
The use of the likelihood function to compare two volatility models will, more likely than not, result in the selection of the worse of the two. Maximum likelihood will lead to an underestimation of the volatility, even with the perfect prediction of the mean. This was found in an analysis of the fidelity of the likelihood function as a means of training and validating a volatility model. A number of cases where the likelihood function leads to an erroneous model are discussed. Systematic errors are corrected by scaling the volatility prediction using a predetermined factor dependent on the number of data points.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 1998
User Contributions:
Comment about this article or add new information about this topic:
A Bayesian analysis of periodic integration
Article Abstract:
Usage of Bayesian techniques revealed the existence of periodic integration in numerous quarterly macroeconomic series in the United Kingdom. The discovery of periodic integration indicated the inappropriateness of traditional unit root models in achieving stochastic trend properties of the series. It also meant the ineffectiveness of standard seasonal adjustment procedures in achieving seasonally corrected time series.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 1997
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: Forecasting growth with time series models. Time series analysis supported by power transformations
- Abstracts: Adjusting judgemental extrapolations using Theil's method and discounted weighted regression. Effect of regressor forecast error on the variance of regression forecasts
- Abstracts: Autoregressive gamma processes. Stochastic models underlying CrostonEs method for intermittent demand forecasting
- Abstracts: Prediction intervals for growth curve forecasts. Order series method for forecasting non-Gaussian time series
- Abstracts: Choosing the number of conditioning events in judgemental forecasting. Going up-going down: how good are people at forecasting trends and changes in trends?