It may be tempting to try to apply standard linear regression methods to the analysis of financial data. However, this may often not work as well as expected. This is because financial data often has more volatility than other sources of data. Whichever statistical model is used, they all have assumptions behind the model that need verification to assess the validity of the model and its results.
When using any model (for instance, Black-Scholes), be sure to know the assumptions behind the model, and to check the assumptions when using the model. If the assumptions behind a statistical model are not met, then this casts doubt on the validity of the model and any conclusions.
A good reference on applying statistical methods to financial data is: Statistics and Finance (Ruppert)
More information is available at https://sites.google.com/a/crlstatistics.net/finecon/
Useful R websites include:
Zurich University of Applied Sciences (ZHAW-IDP) Institute of Data Analysis and Process Design
ETH Zurich Rmetrics
Implementing QuantLib (Luigi Ballabio ebook)
PerformanceAnalytics: Econometric tools for performance and risk analysis
http://cran.r-project.org/web/packages/PerformanceAnalytics
Revolution Webinar: Portfolio Design, Optimization, and Stability Analysis
Presented
Presenter:
Wednesday, January 26, 2011
Diethelm Würtz and Mahendra Mehta, Rmetrics Association
More of interest: