Software engineering is a continually evolving discipline, wherein researchers and members of industry are working towards defining and refining what are known as "best practices." Best practices are the set of known correct engineering techniques that lead to quality software.When a software artifact is produced, it becomes temporally locked into a single instantiation of these best practices at a given point in time. If such software is not maintained in such a way as to keep it current with the evolution of practice, then there is a good chance that subsequent engineers may not understand the design choices made. There are known techniques, called refactorings, which allow for structural changes to software without altering the outward appearance and behavior, thus maintaining the intent of the original design. Unfortunately, refactoring requires an engineer to both understand the techniques to be applied and the code to which they are applied to. This is not always feasible.We have developed an automated system utilizing Evolutionary Algorithms to manipulate refactorings correctly without requiring an underlying understanding of the software. This then allows for sustained levels of quality of evolving software systems. The understandability, maintainability, and reusability of the software regenerate as best practices evolve.

No abstract

In this paper, we present a comparison of five different approaches to extracting decision trees from diagnostic Bayesian nets, including an approach based on the dependency structure of the network itself. With this approach, attributes used in branching the decision tree are selected by a weighted information gain metric computed based upon an associated D-matrix. Using these trees, tests are recommended for setting evidence within the diagnostic Bayesian nets for use in a PHM application. We hypothesized that this approach would yield effective decision trees and test selection and greatly reduce the amount of evidence required for obtaining accurate classification with the associated Bayesian networks. The approach is compared against three alternatives to creating decision trees from probabilistic networks such as ID3 using a dataset forward sampled from the network, KL-divergence, and maximum expected utility. In addition, the effects of using χ2 statistics and probability measures for pre-pruning are examined. The results of our comparison indicate that our approach provides compact decision trees that lead to high accuracy classification with the Bayesian networks when compared to trees of similar size generated by the other methods, thus supporting our hypothesis.

No abstract

No abstract