[1]郭 峰,和萌萌.集成学习融合依赖度的软件缺陷数量预测方法[J].计算机技术与发展,2023,33(06):95-100.[doi:10. 3969 / j. issn. 1673-629X. 2023. 06. 015]
　GUO Feng,HE Meng-meng.Software Defect Quantity Prediction of Ensemble Learning Fusion Dependency[J].,2023,33(06):95-100.[doi:10. 3969 / j. issn. 1673-629X. 2023. 06. 015]

点击复制

集成学习融合依赖度的软件缺陷数量预测方法()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年06期

页码:

95-100

栏目:

软件技术与工程

出版日期:

2023-06-10

文章信息/Info

Title:

Software Defect Quantity Prediction of Ensemble Learning Fusion Dependency

文章编号:

1673-629X(2023)06-0095-06

作者:

郭 峰; 和萌萌

北方工业大学 信息学院,北京 100144

Author(s):

GUO Feng; HE Meng-meng

School of Information Science,North China University of Technology,Beijing 100144,China

关键词:

软件缺陷预测; 依赖度; 集成学习; smote; Stacking

Keywords:

software defect prediction; dependency; ensemble learning; smote; Stacking

分类号:

TP311. 5

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 06. 015

摘要:

软件缺陷可能会导致软件产品故障和经济损失,有效识别潜在的软件缺陷在软件开发、运行维护过程中是至关重要的。 先前软件缺陷预测方面的研究工作主要关注判断一个软件模块是否有

缺陷。 但由于测试资源,预测软件模块缺陷数会对软件开发和维护更有帮助。 因此, 在集成算法基础上, 融合依赖度, 提出一种 新 的 缺 陷 数 预 测 方 法 ELDDP(Ensemble Learning Fusion Dependencies for Software Defect Count Prediction) 。 首先,基于 smote 算法,对从最 k 近邻选择目标类方法进行改进。 将类之间的依赖关系进行度量并引入 smote 算法中,选择目标类时优

先选择依赖度高的类。 其次,提出结合集成学习 Adaboost. R2 算法和 Stacking 模型融合算法构建缺陷数预测模型。 在 Promise 数据集进行实验,对比了三种常用缺陷预测回归模型,实验结果

表明该算法有较好的准确性和稳定性。

Abstract:

Software defects may lead to software product failures and economic losses,and effective identification of potential softwaredefects is critical in the software development,operation and maintenance process. Previous research work in software defect predictionhas focused on determining whether a software module is defective or not. However,due to testing resources,　predicting the number ofsoftware module defects would be more helpful for software development and maintenance. Therefore,a new defect count predictionmethod ELDDP ( Ensemble Learning Fusion Dependencies for Software Defect Count Prediction) is proposed based on the ensemblelearning and fusion dependency. Firstly,the target class selection method from the?

k - most nearest neighbor is improved based on thesmote algorithm. The dependency relationship between classes is measured and introduced into the smote algorithm,and the target classeswith high dependency are selected in preference. Secondly,a combination of ensemble learning Adaboost. R2 algorithm and Stackingmodel fusion algorithm is proposed to construct?

a defect number prediction model. The experiment is carried on Promise dataset,andthree commonly used defect prediction regression models are compared. It is showed that the proposed algorithm has better accuracy andstability.

常用功能