The application of fuzzy enhanced case-based reasoning for identifying fault-prone modules

As highly reliable software is becoming an essential ingredient in many systems, the process of assuring reliability can be a time-consuming, costly process. One way to improve the efficiency of the quality assurance process is to target reliability enhancement activities to those modules that are likely to have the most problems. Within the field of software engineering, much research has been performed to allow developers to identify fault-prone modules within a project. Software quality classification models can select the modules that are the most likely to contain faults so that reliability enhancement activities can be performed to lower the occurrences of software faults and errors. This paper introduces fuzzy logic combined with case-based reasoning (CBR) to determine fault-prone modules given a set of software metrics. Combining these two techniques promises more robust, flexible and accurate models. In this paper, we describe this approach, apply it in a real-world case study and discuss the results. The case study applied this approach to software quality modeling using data from a military command, control and communications (C³) system. The fuzzy CBR model had an overall classification accuracy of more than 85%. This paper also discusses possible improvements and enhancements to the initial model that can be explored in the future