| Software companies possess a lot of data collected during development and maintenance of software systems for the purpose of analysis. Many different techniques and tools have been developed for extracting knowledge. How many and which techniques should be used to analyze a single data set is an important yet difficult question.; In the proposed research the focus is put on a multi-technique approach to analysis of data and extraction of knowledge from it. Models of the same data sets based on Bayesian Belief Network and Decision Tree are built, analyzed and evaluated. Some new methods, evolutionary algorithms, are merged with these techniques for the construction of these models. A number of achievements will be dedicated to comparison between these two approaches in analysis of software engineering data and knowledge extracted for the purposes of defect elimination and complexity perception. |
