Cross-project Software Defect Prediction Research Articles

Feature Selection in Cross-Project Software Defect Prediction

Advances in technology have increased the use and complexity of software. The complexity of the software can increase the possibility of defects. Defective software can cause high losses. Fixing defective software requires a high cost because it can spend up 50% of the project schedule. Most software developers don’t document their work properly so that making it difficult to analyse software development history data. Software metrics which use in cross-project software defects prediction have many features. Software metrics usually consist of various measurement techniques, so there are possibilities for their features to be similar. It is possible that these features are similar or irrelevant so that they can cause a decrease in the performance of classifiers. In this study, several feature selection techniques were proposed to select the relevant features. The classification algorithm used is Naive Bayes. Based on the analysis using ANOVA, the SBS and SBFS models can significantly improve the performance of the Naïve Bayes model.

Within-project and cross-project software defect prediction based on improved transfer Naive Bayes algorithm

With the continuous expansion of software scale, software update and maintenance have become more and more important. However, frequent software code updates will make the software more likely to introduce new defects. So how to predict the defects quickly and accurately on the software change has become an important problem for software developers. Current defect prediction methods often cannot reflect the feature information of the defect comprehensively, and the detection effect is not ideal enough. Therefore, we propose a novel defect prediction model named ITNB (Improved Transfer Naive Bayes) based on improved transfer Naive Bayesian algorithm in this paper, which mainly considers the following two aspects: (1) Considering that the edge data of the test set may affect the similarity calculation and final prediction result, we remove the edge data of the test set when calculating the data similarity between the training set and the test set; (2) Considering that each feature dimension has different effects on defect prediction, we construct the calculation formula of training data weight based on feature dimension weight and data gravity, and then calculate the prior probability and the conditional probability of training data from the weight information, so as to construct the weighted bayesian classifier for software defect prediction. To evaluate the performance of the ITNB model, we use six datasets from large open source projects, namely Bugzilla, Columba, Mozilla, JDT, Platform and PostgreSQL. We compare the ITNB model with the transfer Naive Bayesian (TNB) model. The experimental results show that our ITNB model can achieve better results than the TNB model in terms of accurary, precision and pd for within-project and cross-project defect prediction.

Dimensional Reduction on Cross Project Defect Prediction

The complexity of the software can increase the possibility of defects. Defective software can cause high losses. The software containing defects can cause large losses. Most software developers don’t document their work properly so that making it difficult to analyse software development history data. The cross-project software defect prediction used several datasets from different projects and combining for training and testing. The dataset with high dimension can cause bias, contain irrelevance data, and require large resources to process it. In this study, several dimensional reduction algorithm and Decision Tree as classifier. Based on the analysis using ANOVA, all models that implement dimensional reduction can significantly improve the performance of the Decision Tree model.

Tackling Imbalanced Class on Cross-Project Defect Prediction Using Ensemble SMOTE

The dataset with imbalanced class can reduce the performance of the classifiers. In this study proposed a cross-project software defect prediction model that applies the SMOTE (Synthetic Minority Oversampling Technique) to balance classes in datasets and ensembles technique to reduce misclassification. The ensemble technique using AdaBoost and Bagging algorithms. The results of the study show that the model that integrates SMOTE and Bagging provides better performance. The proposed model can find more software defects and more precise.

Ensemble Undersampling to Handle Unbalanced Class on Cross-Project Defect Prediction

There has been much research which proposed for cross-project software defect prediction models but no models that perform very well with various datasets in general. Software defect dataset usually imbalanced because it contains far more the not defected modules than the defected modules. Class imbalances in the dataset can reduce the performance of classifiers in the software defect prediction model. In this study proposed a Random Undersampling algorithm to balance classes and ensemble techniques to reduce misclassification. The ensemble technique used is the AdaBoost and Bagging algorithm. The results showed that the software defect prediction model that integrates the Random Undersampling algorithm and AdaBoost provides better performance and can find more defects than other models.

The cross-project defect prediction based on PSO and Feature Dependent Naive Bayes

The software defect prediction often assumes that the software under test has rich historical information, which is a harsh condition in reality. Cross-project defect prediction can predict projects with rare historical information by using historically informative projects. The differences in software architecture and code environment pose challenges to cross-project software defect prediction. Based on the characteristics of software defect prediction sets, this paper introduces a two-stage algorithm based on particle swarm optimization algorithm and Feature Dependent Naive Bayesian classifier.

A Cluster Based Feature Selection Method for Cross-Project Software Defect Prediction

Cross-project defect prediction (CPDP) uses the labeled data from external source software projects to compensate the shortage of useful data in the target project, in order to build a meaningful classification model. However, the distribution gap between software features extracted from the source and the target projects may be too large to make the mixed data useful for training. In this paper, we propose a cluster-based novel method FeSCH (Feature Selection Using Clusters of Hybrid-Data) to alleviate the distribution differences by feature selection. FeSCH includes two phases. The feature clustering phase clusters features using a density-based clustering method, and the feature selection phase selects features from each cluster using a ranking strategy. For CPDP, we design three different heuristic ranking strategies in the second phase. To investigate the prediction performance of FeSCH, we design experiments based on real-world software projects, and study the effects of design options in FeSCH (such as ranking strategy, feature selection ratio, and classifiers). The experimental results prove the effectiveness of FeSCH. Firstly, compared with the state-of-the-art baseline methods, FeSCH achieves better performance and its performance is less affected by the classifiers used. Secondly, FeSCH enhances the performance by effectively selecting features across feature categories, and provides guidelines for selecting useful features for defect prediction.