Mining metrics to predict component failures

Nagappan, Nachiappan; Ball, Thomas; Zeller, Andreas

doi:10.1145/1134285.1134349

Cited by 622 publications

(413 citation statements)

References 20 publications

Supporting

Mentioning

383

Contrasting

Unclassified

Order By: Relevance

“…Approaches range from static code analysis and mining of software repositories and bug databases [8,9,10] to dynamic program verification. The latter focus on the data flow [11,12] or, like all call graph based techniques, on the control flow [13,14].…”

Section: Related Workmentioning

confidence: 99%

Mining Edge-Weighted Call Graphs to Localise Software Bugs

Eichinger¹,

Böhm²,

Huber³

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

An important problem in software engineering is the automated discovery of noncrashing occasional bugs. In this work we address this problem and show that mining of weighted call graphs of program executions is a promising technique. We mine weighted graphs with a combination of structural and numerical techniques. More specifically, we propose a novel reduction technique for call graphs which introduces edge weights. Then we present an analysis technique for such weighted call graphs based on graph mining and on traditional feature selection schemes. The technique generalises previous graph mining approaches as it allows for an analysis of weights. Our evaluation shows that our approach finds bugs which previous approaches cannot detect so far. Our technique also doubles the precision of finding bugs which existing techniques can already localise in principle.

show abstract

Section: Related Workmentioning

confidence: 99%

Mining Edge-Weighted Call Graphs to Localise Software Bugs

Eichinger¹,

Böhm²,

Huber³

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

show abstract

“…Nagappan et al [31] used catalog of source code metrics to predict post release defects at module level on five Microsoft systems and found it was possible to build predictors for one individual project, but that no predictor would perform well on all projects.…”

Section: Related Workmentioning

confidence: 99%

Empirical Evaluation of Machine Learning Algorithms for Fault Prediction

Kaur¹,

Kaur²

2014

LNSE

View full text Add to dashboard Cite

Abstract-Producing quality software is a very challenging task looking at the size and complexity of software developed these days. Predicting software quality early helps in using testing resources optimally. So, many statistical and machine learning techniques are used to predict quality classes in software. In this work, six machine learning classifiers have been used to estimate the fault proneness of 5885 classes used in five open source software on the basis of object-oriented metrics calculated on these classes. Bagging and J48 classifiers turn out to be the best one amongst the classifiers used.

show abstract

“…Dagpinar et al [15] investigate the significance of different OO metrics for the purpose of predicting maintainability of software. Nagappan et al [25] focus on mining metrics to predict component failures. The authors noted that there is no a single set of complexity metrics that could be used as a universally best defect predictor.…”

Section: Related Workmentioning

confidence: 99%

Metrics and Software Quality Evolution: A Case Study on Open Source Software

Drouin¹,

Badri²,

Touré³

2013

IJCTE

View full text Add to dashboard Cite

Abstract-This paper aims at analyzing empirically the quality evolution of an open source software using metrics. We used a control flow based metric (Quality Assurance IndicatorQi) which we proposed in a previous work. We wanted to investigate if the Qi metric can be used to observe how quality evolves along the evolution of the successive released versions of the subject software system. We addressed software quality from an internal perspective. We performed an empirical analysis using historical data on the subject system (Apache Tomcat). The collected data cover, in fact, a period of more than seven years (thirty-one versions in total). Empirical results provide evidence that the Qi metric reflects properly the quality evolution of the subject system.

show abstract

Mining metrics to predict component failures

Cited by 622 publications

References 20 publications

Mining Edge-Weighted Call Graphs to Localise Software Bugs

Mining Edge-Weighted Call Graphs to Localise Software Bugs

Empirical Evaluation of Machine Learning Algorithms for Fault Prediction

Metrics and Software Quality Evolution: A Case Study on Open Source Software

Contact Info

Product

Resources

About