An improved transfer adaptive boosting approach for mixed‐project defect prediction

Gong, Lina; Jiang, Shujuan; Jiang, Li

doi:10.1002/smr.2172

Cited by 14 publications

(12 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For different prediction tasks, SDP includes binary classification, 17,18,21,24,29–32 numeric, 1,33,34 ranking, 35,36 and association rule mining 37,38 . According to different sources of training data, SDP can be divided into three scenarios: WPDP, CPDP, and mixed‐project defect prediction (MPDP) 10,39,40 …”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

WIFLF: An approach independent of the target project for cross‐project defect prediction

Cui

Liu

Wang

2022

J Software Evolu Process

View full text Add to dashboard Cite

Cross-project defect prediction (CPDP) is used to build defect prediction models when data from the target project are not enough. There has been several approaches to improve the performance of CPDP, such as feature transformation and instance selection methods. However, existing techniques are strongly dependent on the target data to reduce the distribution discrepancy between source and target projects. That is, the performance of these methods is determined by the effectiveness of feature transformation or the similarity between two projects. Additionally, when there is a large amount of source data that needs to be matched with target data, it will take much time and reduce the efficiency of model construction.Therefore, it is vital to explore a target project-agnostic approach to build CPDP models. This paper presents a Weighted Isolation Forest with class Label information Filter (WIFLF) to relieve the issues above. Four groups of datasets from AEEEM, Relink and PROMISE Data Repository are used to conduct CPDP models. Besides, WIFLF is compared with 12 approaches. The experimental results indicate that WIFLF significantly outperforms all the baselines. Specifically, WIFLF with random forest significantly improves the performance over the baselines on average by at least 14.64% and 4.90% with respect to Skewed F-Measure and G-Measure, respectively.

show abstract

Section: Related Workmentioning

confidence: 99%

“…37,38 According to different sources of training data, SDP can be divided into three scenarios: WPDP, CPDP, and mixed-project defect prediction (MPDP). 10,39,40 For the WPDP scenario, source data named within-project (WP) data are from the previous version of the target data or the same project.…”

Section: Related Workmentioning

confidence: 99%

WIFLF: An approach independent of the target project for cross‐project defect prediction

Cui

Liu

Wang

2022

J Software Evolu Process

View full text Add to dashboard Cite

show abstract

“…The classic instance-based transfer learning method is TrAdaBoost [29], which uses AdaBoost-based technology to reweight the instances in the source domain and transfer the useful instances with heavy weights in the source domain to the target domain. There are many improved methods based on TrAdaBoost, such as DTrBoost [26], ITrAdaBoost [27], TrResampling [28], TaskTrAdaBoost [30], and Trans-Boost [31]. The structure of instance-based transfer learning is as follows:…”

Section: B Instance-based Transfer Learningmentioning

confidence: 99%

“…It is well known that an ensemble of classifiers is usually significantly more accurate than a single learner, and ensemble methods have already achieved great success in many realworld tasks [22] [23]. The AdaBoost algorithm [24] is the most influential boosting algorithm [25], it has combined with many transfer learning methods for classification, such as DTrBoost [26], ITrAdaBoost [27], TrResampling [28], TrAdaBoost [29], TaskTrAdaBoost [30], and Trans-Boost [31], all of them use AdaBoost-based technology to re-weight the instances in the source domain and transfer useful instances with the heaviest weights in the source domain to the target domain. To boost the performance of deep learning-based HSI classification, Chen et al [32] proposed two deep learning ensemble-based classification methods: CNNs and deep residual networks are used as individual classifiers and random subspaces, and then the learned weights are transferred from one individual classifier to another.…”

Section: Introductionmentioning

confidence: 99%

Extreme Learning Machine-Based Ensemble Transfer Learning for Hyperspectral Image Classification

Liu

Cai

et al. 2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Although the extreme learning machine (ELM) has been successfully applied to hyperspectral image (HSI) classification, the development of the ELM is restricted by insufficient training data. In this paper, we propose a novel extreme learning machine-based ensemble transfer learning algorithm for hyperspectral image classification named TL-ELM. TL-ELM not only retains the input weights and hidden biases of the ELM learned from the target domain, but also utilizes instances in the source domain to iteratively adjust the output weights of the ELM, which are used as the weights of the training models, and then ensembles the training models with their weights for the final classification. In experiments, we choose different regions in northern Italy, namely, Pavia University and Pavia Centre, as the source data set and target data set, respectively, and through a comparison with other transfer learning algorithms, we demonstrate that our proposed TL-ELM algorithm is superior on HSI classification tasks with only a few labeled data points in the target domain. Furthermore, we set Pavia University as the source data set and Pavia Centre as the target data set to demonstrate that our proposed method can effectively transfer useful instances between different HSIs.

show abstract

“…The National Institute of Standards and Technology (NIST) estimates that software defect costs the United States approximately $60 billion annually, and identifying and fixing defects could save $22 billion (Cai et al, 2019). Software defect prediction helps optimize resource allocation and detect defects in time (Gong et al, 2019). Due to the drawbacks of traditional software defect prediction such as non-immediacy and coarse granularity, researchers have proposed a change-level software defect prediction, which is also known as just-in-time software defect prediction (JIT-SDP) (Kamei et al, 2012).…”

Section: Introductionmentioning

confidence: 99%

EANDL: A High-performance and Low-cost Effort-Aware Just-in-Time Software Defect Prediction Model

Chen,

Wang,

Song

et al. 2023

Preprint

View full text Add to dashboard Cite

Just-in-time software defect prediction (JIT-SDP) is a fine-grained software defect prediction technology and its predictive entity is code changes. Effort-aware just-in-time software defect prediction (Effort-aware JIT-SDP) is a defect prediction technology that takes the cost of detection into consideration with the goal of detecting more defect changes with less cost. To solve the problem of poor prediction performance of Effort aware JIT-SDP model, an Effort-aware JIT-SDP model called EANDL based on NSGA3, Differential Evolution and Logistic Regression is proposed. Firstly, two evaluation metrics called DPR and LDSOD are proposed to evaluate the sorting performance. Secondly, the prediction model is constructed by using logistic regression and maximizing two evaluation metrics using NSGA3-DE so as to determine the optimal coefficient vector of logistic regression. In addition, two metrics called ACE and ACEP to measure the effort consumption of Effort-aware JIT-SDP models are proposed. Finally, intensive simulation experiments have been conducted on six open-source datasets to demonstrate the performance of the proposed method. Empirical results demonstrate that EANDL outperforms eight typical and classical supervised and unsupervised models in terms of sorting performance, effort consumption, and response time.

show abstract

An improved transfer adaptive boosting approach for mixed‐project defect prediction

Cited by 14 publications

References 46 publications

WIFLF: An approach independent of the target project for cross‐project defect prediction

WIFLF: An approach independent of the target project for cross‐project defect prediction

Extreme Learning Machine-Based Ensemble Transfer Learning for Hyperspectral Image Classification

EANDL: A High-performance and Low-cost Effort-Aware Just-in-Time Software Defect Prediction Model

Contact Info

Product

Resources

About