Bayesian Logistic Regression for software defect prediction (S)

Software defects are one of the main contributors to information technology waste and lead to rework, thus consuming a lot of time and money. Software defect prediction has the objective of defect prevention by classifying certain modules as defective or not defective. Many researchers have conducted research in the field of software defect prediction using NASA MDP public datasets, but these datasets still have shortcomings such as class imbalance and noise attribute. The class imbalance problem can be overcome by utilizing SMOTE (Synthetic Minority Over-sampling Technique) and the noise attribute problem can be solved by selecting features using Particle Swarm Optimization (PSO), So in this research, the integration between SMOTE and PSO is applied to the classification technique machine learning naïve Bayes and logistic regression. From the results of experiments that have been carried out on 8 NASA MDP datasets by dividing the dataset into training and testing data, it is found that the SMOTE + PSO integration in each classification technique can improve classification performance with the highest AUC (Area Under Curve) value on average 0,89 on logistic regression and 0,86 in naïve Bayes in the training and at the same time better than without combining the two.

Download Full-text

A Novel Feature Selection Method Based on Maximum Likelihood Logistic Regression for Imbalanced Learning in Software Defect Prediction

The International Arab Journal of Information Technology ◽

10.34028/iajit/17/5/5 ◽

2020 ◽

Vol 17 (5) ◽

pp. 721-730

Author(s):

Kamal Bashir ◽

Tianrui Li ◽

Mahama Yahaya

Keyword(s):

Machine Learning ◽

Logistic Regression ◽

Feature Selection ◽

Maximum Likelihood ◽

Defect Prediction ◽

Feature Subset ◽

Software Defect Prediction ◽

Software Defect ◽

Optimal Feature Subset ◽

Optimal Feature

The most frequently used machine learning feature ranking approaches failed to present optimal feature subset for accurate prediction of defective software modules in out-of-sample data. Machine learning Feature Selection (FS) algorithms such as Chi-Square (CS), Information Gain (IG), Gain Ratio (GR), RelieF (RF) and Symmetric Uncertainty (SU) perform relatively poor at prediction, even after balancing class distribution in the training data. In this study, we propose a novel FS method based on the Maximum Likelihood Logistic Regression (MLLR). We apply this method on six software defect datasets in their sampled and unsampled forms to select useful features for classification in the context of Software Defect Prediction (SDP). The Support Vector Machine (SVM) and Random Forest (RaF) classifiers are applied on the FS subsets that are based on sampled and unsampled datasets. The performance of the models captured using Area Ander Receiver Operating Characteristics Curve (AUC) metrics are compared for all FS methods considered. The Analysis Of Variance (ANOVA) F-test results validate the superiority of the proposed method over all the FS techniques, both in sampled and unsampled data. The results confirm that the MLLR can be useful in selecting optimal feature subset for more accurate prediction of defective modules in software development process

Download Full-text