基于PCA-Smote-XGBoost的软件缺陷预测研究  

Research on Software Defect Prediction Based on PCA-Smote-XGBoost

作　　者：曾子安 李英梅[1] 

机构地区：[1]哈尔滨师范大学计算机科学与信息工程学院,黑龙江 哈尔滨

出　　处：《软件工程与应用》2024年第3期346-357,共12页Software Engineering and Applications

摘　　要：随着软件系统的复杂性日益增加,软件缺陷预测成为了确保软件质量的重要手段。本研究提出了一种基于PCA-Smote-XGBoost的软件缺陷预测模型,旨在提高缺陷预测的准确性和效率。本文采用主成分分析(PCA)进行数据降维,保留95%的方差,以减少特征数量并提取关键信息;利用Smote过采样方法解决数据不平衡问题;结合XGBoost算法构建预测模型,并通过实验验证模型的有效性。在软件缺陷预测常用数据集的十一个项目中,实验结果表明,该模型在软件缺陷预测方面相较于其他八种基准模型,具有最高的准确率ACC和F1,能够有效地辅助软件开发团队识别潜在的缺陷风险。With the increasing complexity of software systems, software defect prediction has become an important means to ensure software quality. This study proposes a software defect prediction model based on PCA-Smote-XGBoost, aiming to improve the accuracy and efficiency of defect prediction. This article uses Principal Component Analysis (PCA) for data dimensionality reduction, retaining 95% of the variance to reduce the number of features and extract key information;uses Smote oversampling method to solve the problem of data imbalance;builds a prediction model using the XGBoost algorithm and validates its effectiveness through experiments. Among the eleven commonly used datasets for software defect prediction, experimental results show that the model has the highest accuracy ACC and F1 compared to the other eight benchmark models in software defect prediction, and can effectively assist software development teams in identifying potential defect risks.

关 键 词：软件缺陷预测 PCA SMOTE XGBoost 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]