Paper Info
Title
Comparative analysis of statistical and machine learning methods for predicting faulty modules.
Abstract
•We investigate the performance of the fault proneness predictions using static code metrics.•The performance of logistic regression and six machine-learning methods (ANN, SVM, DT, CCN, GMDH and GEP) was evaluated in the study for predicting the fault proneness in the modules.•Thus, the aim of this work is to find the best predictor for predicting faulty modules.•The validation of the methods is carried out using Receiver Operating Characteristic (ROC) analysis.•In order to perform the analysis we validate the performance of these methods using public domain AR1 and AR6 data sets.
Year
DOI
Venue
2014
10.1016/j.asoc.2014.03.032
Applied Soft Computing
Keywords
Field
DocType
Software quality,Static code metrics,Logistic regression,Machine learning,Receiver Operating Characteristic (ROC) curve
Data mining,Online machine learning,Decision tree,Computer science,Support vector machine,Software,Artificial intelligence,Software quality,Artificial neural network,Group method of data handling,Machine learning,Software development
Journal
Volume
ISSN
Citations 
21
1568-4946
19
PageRank 
References 
Authors
0.78
29
1
Name
Order
Citations
PageRank
Ruchika Malhotra153335.12