Deriving Numerical Abstract Domains via Principal Component Analysis

Numerical Abstract Domains via Principal Component Analysis Gianluca Amato, Maurizio Parton, and Francesca Scozzari Universit`a di Chieti-Pescara – Dipartimento di Scienze We propose a new technique for developing ad-hoc numerical domains by means of statistical analysis. We apply Principal Component Analysis to partial execution traces of programs, to find out a “best basis” in the vector space of program variables. This basis may to specialize numerical abstract domains, in order to enhance the precision of the analysis. As an example, we apply our technique to interval analysis of simple imperative programs.

• Type: Journal
• Year: 2010
• Where: SAS
• Authors: Gianluca Amato, Maurizio Parton, Francesca Scozzari
• Keywords: Ad-hoc Numerical Domains, Formal Methods, Numerical Abstract Domains, Principal Component Analysis, SAS 2010