Comparing design and code metrics for software quality prediction

Abstract

The prediction of fault-prone modules continues to attract interest due to the significant impact it has on software quality assurance. One of the most important goals of such techniques is to accurately predict the modules where faults are likely to hide as early as possible in the development lifecycle. Design, code, and most recently, requirements metrics have been successfully used for predicting fault-prone modules. The goal of this paper is to compare the performance of predictive models which use design-level metrics with those that use code-level metrics and those that use both. We analyze thirteen datasets from NASA Metrics Data Program which offer design as well as code metrics. Using a range of modeling techniques and statistical significance tests, we confirmed that models built from code metrics typically outperform design metrics based models. However, both types of models prove to be useful as they can be constructed in different project phases. Code-based models can be used to increase the performance of design-level models and, thus, increase the efficiency of assigning verification and validation activities late in the development lifecycle. We also conclude that models that utilize a combination of design and code level metrics outperform models which use either one or the other metric set.

Integrating software quality models into risk-based testing

“…Nagappan et al (2006) found that code complexity metrics are an effective mean to predict defects in software code. Further research showed (i.e., Catal et al (2007), Jiang et al (2008), Radjenovic et al (2013)) that software metrics in general are useful for predicting defects and their location in software products. In the following, Felderer et al (2012, p. 163) claim that the employment of different metrics, which can usually be determined automatically, can serve as a basis for determining the probability factor.…”

Predicting Faults before Testing Phase using Halstead's Metrics

“…al., compare the performance of predictive models which use design level metrics with those uses code level metrics and those that use both. [14] Nick J. Pizzi purposes in his case study an aggregation technique based on fuzzy integration that combines that combines the predictive quantitative assessments from multiple classifier [15] Cagatay Catal surveys the software engineering literature on software fault prediction and both machine learning based and statistical based approaches on 90 software fault prediction papers [16] YANG Weimin, LI Longshu purposes the correlation of software metrics focusing on the data sets of software defect prediction. A rough set model is presented to reduce the attributes of data sets of software defect prediction [17].…”

Using Data Mining for Static Code Analysis of C

“…Another contribution to the area by Jiang et.al. [13] who compared the performance of design and code metrics in fault prediction models. They came to the conclusion that models using a combination of these metric outperform models that use either code or design metrics alone.…”