Towards Improving Software Architecture Degradation Mitigation by Machine Learning
Year: 2020
Type: conference-paper
Abstract: Mitigating software architecture degradation is a task in evolving large and complex software-intensive systems that is as important as it is challenging. One aspect adding to the complexity of the task is the amount of information in the implementations of most real-world systems to be digested in order to detect, analyse, and remedy degradation. In domains with similar challenges, machine learning techniques have been applied in recent years and partially delivered exciting results. Hence the question arises whether, and to which degree, machine learning can be successfully applied to tackle software architecture degradation. In this paper, we propose a novel combination of existing techniques for different phases of the task of mitigating software architecture degradation from detecting it to repairing it. We outline how these techniques could be complemented by machine learning to increase their accuracy and efficiency over time
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3136382286