Bayesian reasoning for software testing by Akbar Siami Namin, Mohan Sridharan

Abstract:

Despite significant advances in software testing research, the ability to produce reliable software products for a variety of critical applications remains an open problem. The key challenge has been the fact that each program or software product is unique, and existing methods are predominantly not capable of adapting to the observations made during program analysis. This paper makes the following claim: Bayesian reasoning methods provide an ideal research paradigm for achieving reliable and efficient software testing and program analysis. A brief overview of some popular Bayesian reasoning methods is provided, along with a justification of why they are applicable to software testing. Furthermore, some practical challenges to the widespread use of Bayesian methods are discussed, along with possible solutions to these challenges.

http://portal.acm.org/citation.cfm?id=1882433

By in Business on May 20, 2011