Shotgun correlations in software measures

Courtney, Richard E.; Gustafson, David A.

doi:10.1049/sej.1993.0002

Cited by 48 publications

(26 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Correspondingly, the null hypothesis (H0) is that there is no correlation between such. We choose a correlation test based on insights from Bagheri et al [5] and Courtney et al [17]. The former observe that their structural metrics did not have a normal distribution.…”

Section: Methodsmentioning

confidence: 99%

“…We are aware that our sample size (eight systems) is not sufficient and may give rise to unreliable statistics [17]. We mitigate this threat by choosing a non-parametric measure (as suggested by [17]), and by a qualitative inspection of strong (and significant) correlations based on our knowledge of the languages.…”

Section: Threats To Validitymentioning

confidence: 99%

“…We mitigate this threat by choosing a non-parametric measure (as suggested by [17]), and by a qualitative inspection of strong (and significant) correlations based on our knowledge of the languages.…”

Section: Threats To Validitymentioning

confidence: 99%

See 2 more Smart Citations

Towards system analysis with variability model metrics

Berger

Guo

2014

Proceedings of the Eighth International Workshop on Variability Modelling of Software-Intensive Systems

View full text Add to dashboard Cite

Variability models are central artifacts in highly configurable systems. They aim at planning, developing, and configuring systems by describing configuration knowledge at different levels of formality. The existence of large models using a variety of modeling concepts in heterogeneous languages with intricate semantics calls for a unified measuring approach. In this position paper, we attempt to take a first step towards such a measurement. We discuss perspectives of metrics, define low-level measurement goals, and conceive and implement metrics based on variability modeling concepts found in real-world languages and models. An evaluation of these metrics with real-world models and codebases provides insight into the benefits of such metrics for the defined perspectives.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Threats To Validitymentioning

confidence: 99%

See 1 more Smart Citation

Towards system analysis with variability model metrics

Berger

Guo

2014

Proceedings of the Eighth International Workshop on Variability Modelling of Software-Intensive Systems

View full text Add to dashboard Cite

show abstract

“…There are several ways to decide if two sets of values are close. We could determine the error between the two sets or compute correlation using Pearson's correlation equation [7]. We use accuracy (defined in Equation 5) between a single similarity factor and the OracleSet to decide weights of each factor.…”

Section: Deciding Factor Weightsmentioning

confidence: 99%

When functions change their names: automatic detection of origin relationships

Kim

Pan

Whitehead

2005

12th Working Conference on Reverse Engineering (WCRE'05)

View full text Add to dashboard Cite

show abstract

“…This has been referred to as the shotgun approach 7 or more generally, data mining 8 . Courtney and Gustafson 7 also discuss the dangers of relying on correlation coefficients where hypotheses have not been proposed in advance.…”

Section: Least Squares Regressionmentioning

confidence: 99%