Predicting defect densities in source code files with decision tree learners

Knab, Patrick; Pinzger, Martin; Bernstein, Abraham

doi:10.1145/1137983.1138012

Cited by 98 publications

(61 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Results clearly underline that defect prediction models have to take into account different aspects and measures of the software development and maintenance [1]. In extension to our previous work on predicting defect density of source files [2] we use detailed evolution data from an industrial software project and include team structure and process measures.…”

Section: Introductionmentioning

confidence: 92%

“…To build a balanced prediction model we create features to represent several important aspects of software development such as the complexity of the designed solution, process used for development, interrelation of classes, etc. As previous studies [2,3] discovered that relative features provide better performance in prediction than absolute ones, we decided that all our 63 features have to be relative. For EQ-Mina we set up the following categories of features for each file containing changes within the inspection period:…”

Section: Featuresmentioning

confidence: 99%

See 1 more Smart Citation

EQ-Mine: Predicting Short-Term Defects for Software Evolution

Ratzinger

Pinzger

Gall

Fundamental Approaches to Software Engineering

Self Cite

View full text Add to dashboard Cite

Abstract. We use 63 features extracted from sources such as versioning and issue tracking systems to predict defects in short time frames of two months. Our multivariate approach covers aspects of software projects such as size, team structure, process orientation, complexity of existing solution, difficulty of problem, coupling aspects, time constrains, and testing data. We investigate the predictability of several severities of defects in software projects. Are defects with high severity difficult to predict? Are prediction models for defects that are discovered by internal staff similar to models for defects reported from the field?We present both an exact numerical prediction of future defect numbers based on regression models as well as a classification of software components as defect-prone based on the C4.5 decision tree. We create models to accurately predict short-term defects in a study of 5 applications composed of more than 8.000 classes and 700.000 lines of code. The model quality is assessed based on 10-fold cross validation.

show abstract

Section: Introductionmentioning

confidence: 92%

Section: Featuresmentioning

confidence: 99%

EQ-Mine: Predicting Short-Term Defects for Software Evolution

Ratzinger

Pinzger

Gall

Fundamental Approaches to Software Engineering

Self Cite

View full text Add to dashboard Cite

show abstract

“…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%

“…Using standard complexity measures from software engineering, the source code is mined with regression models, which can then predict post-release failures for new software entities. A similar study uses decision trees to predict failure probabilities [9]. [10] uses regression techniques to predict the likelihood of bugs based on static usage relationships between software components.…”

Section: Mining Software Metrics and Invariantsmentioning

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

“…Knab and Pinzger [15] applied a decision tree-based algorithm to predict defect densities in source code files. They extracted modifications, defect report metrics, number of incoming and outgoing calls from source releases and version history database.…”

Section: Related Workmentioning

confidence: 99%

A Bayesian Network Based Approach for Change Coupling Prediction

Zhou

Würsch

Giger

et al. 2008

2008 15th Working Conference on Reverse Engineering

View full text Add to dashboard Cite

Source code coupling and change history are two important data sources for change coupling analysis. The popularity of public open source projects in recent years makes both sources available. Based on our previous research, in this paper, we inspect different dimensions of software changes including change significance or source code dependency levels, extract a set of features from the two sources and propose a bayesian network-based approach for change coupling prediction. By combining the features from the co-changed entities and their dependency relation, the approach can model the underlying uncertainty. The empirical case study on two medium-sized open source projects demonstrates the feasibility and effectiveness of our approach compared to previous work. A Bayesian Network Based Approach for Change Coupling Prediction AbstractSource code coupling and change history are two important data sources for change coupling analysis. The popularity of public open source projects in recent years makes both sources available. Based on our previous research, in this paper, we inspect different dimensions of software changes including change significance or source code dependency levels, extract a set of features from the two sources and propose a bayesian network-based approach for change coupling prediction. By combining the features from the co-changed entities and their dependency relation, the approach can model the underlying uncertainty. The empirical case study on two medium-sized open source projects demonstrates the feasibility and effectiveness of our approach compared to previous work.

show abstract

Predicting defect densities in source code files with decision tree learners

Cited by 98 publications

References 14 publications

EQ-Mine: Predicting Short-Term Defects for Software Evolution

EQ-Mine: Predicting Short-Term Defects for Software Evolution

Mining Edge-Weighted Call Graphs to Localise Software Bugs

A Bayesian Network Based Approach for Change Coupling Prediction

Contact Info

Product

Resources

About