Machine learning for predicting defect-inducing software changes

Software systems have become ever larger and more complex. This inevitably leads to software defects, whose debugging is estimated to cost the global economy 312 billion USD annually. Reducing the number of software defects is challenging, and particularly important given the strong pressure towards rapid delivery. Such pressure impedes different parts of the software source code to all receive equally large amount of inspection and testing effort.

In order to help software engineers to target special testing and inspection attention towards parts of the source code most likely to induce defects, this project will develop novel machine learning approaches for predicting defect-inducing changes in the source code. These approaches will use data produced during the software development process to create machine learning models able to provide such predictions as soon as software changes are implemented. It is expected that, by using these approaches, the risk of software developers committing defect-inducing software changes will be reduced, saving debugging cost and ultimately increasing software quality.

The student will start the project by reviewing existing literature on machine learning and prediction of defect-inducing software changes. Novel machine learning approaches will then be developed to overcome problems of existing approaches. The new approaches will be evaluated and compared against other approaches based on existing software projects.