Resource-oriented software quality classification models

Developing high-quality software within the allotted time and budget is a key element for a productive and successful software project. Software quality classification models that provide a risk-based quality estimation, such as fault-prone ( fp) and not fault-prone ( nfp), have proven their usefulness as software quality assurance techniques. However, their usefulness is largely dependent on the availability of resources for deploying quality improvements to modules predicted as fp. Since every project has its own special needs and specifications, we feel a classification modeling approach based on resource availability is greatly warranted. We propose and demonstrate the use of a resource-based measure, i.e., “Modified Expected Cost of Misclassification” ( mecm), for selecting and evaluating classification models. It is an extension of the “Expected Cost of Misclassification” ( ecm) measure, which we have previously applied for model-evaluation purposes. The proposed measure facilitates building resource-oriented classification models and overcomes the limitation of ecm, which assumes that enough resources are available to enhance all modules predicted as fp. The primary aspect of mecm is that it penalizes a model, in terms of costs of misclassifications, if the model predicts more number of fp modules than the number that can be enhanced with the available resources. Based on the resources available for improving quality of software modules, a practitioner can use the proposed methodology to select a model that best-suits the projects goals. Hence, the best possible and practical usage of the available resources can be achieved. The application, analysis, and benefits of mecm is shown by developing models using Logistic Regression. It is concluded that the use of mecm is a promising approach for practical software quality improvement.