改进PSO与模糊积分软件缺陷预测方法研究
|
摘要
随着计算机的广泛应用,计算机软件的需求量逐渐增大,如何高效开发高质量的计算机软件成为软件公司关注的问题。计算机软件的早期开发理念和方法在很大程度上限制了计算机的开发质量和开发周期。软件缺陷预测是软件工程领域中为了提高软件开发效率,降低开发成本,同时又能提高软件的开发质量所作的研究。该方法通过研究软件中的高缺陷模块,运用统计或分类方法预测软件模块含有缺陷的情况。软件开发人员在开发过程中将预测的结果作为参考来关注易于产生缺陷的模块,从而提高软件的开发质量,降低软件的开发周期。
目前,软件缺陷预测技术的研究成为软件工程和应用中重要领域之一,该项技术的研究也越来越多。针对软件缺陷预测方法对提高软件开发质量和软件开发效率具有重要的指导意义,本文首先对计算机软件缺陷预测的基本概念、意义、研究现状进行了详细的描述,在对现有的缺陷预测技术方法进行了总结概括的基础上,分析比较了已有软件缺陷预测的优点及不足,重点研究了模糊积分理论,根据模糊积分的映射功能设计了缺陷预测模型。该模型在高维空间对软件模块数据分类并建立分类超平面完成对模块的预测。针对软件中缺陷模块和非缺陷模块的比例相差较大问题采用错误分类惩罚机制调整最小错误分类率的模糊积分分类函数。最优化分类超平面函数参数优化根据遗传优化方法来实现。其次,详细分析了软件模块属性特征与软件缺陷的关系,运用改进粒子群优化算法建立软件缺陷的预测模型。该模型通过优化提取缺陷模块特征属性规则进行预测。实验结果表明了所设计模型的有效性和适用性。
针对模糊测度和模糊积分所具有的特性,本文还定义了在粒子群优化过程中粒子个体间的交互作用和在迭代过程中惯性权重的综合运用。在迭代过程中,粒子的搜索方向及速度随着所有粒子之间关系的变化而变化,惯性权重的改变是通过综合迭代过程中的原有的惯性权重所占比重自动修改。实验结果表明基于模糊积分的粒子群优化算法是可行的。
Along with the extensive application of computers, the demand for computer software to grow. How to develop high-quality software becomes the key question that software company pay close attention to. But, the early concepts and methods of software development have limited the development of quality and development cycle of the computer in a large extent. Software defect prediction is the research that to improve software development efficiency, reduce development costs and improve the quality of software development in the field of software engineering. Study of the defect modules of software, predict erroneous software modules by classify and statistical analysis methods. The predict result is used as a reference, software developers will focus quality assurance activities on defect-prone modules and thus improve software quality and shortening development cycle by using resources more efficiently in the development process.
Now, software defect predict research is becoming an important field in software engineering and applications. And many researchers have already focused on it. Software defect prediction method has great consultive value and important significance to improving software quality and efficiency of software development. The paper begins with an elaborate of basic conceptions, significance and research status about software defect predict based on the survey of the key predict method, and analyzed the disadvantages of existing prediction methods, then researches the fuzzy integral theory, we put forward a software defect predict model based on the mapping of fuzzy integral. It adopts method with ability of clusters from data which has high dimensions to establish the classification hyper-plane in high dimensional space. And draw into the minimum ratio of misclassification to improve the classification. Simulated genetic algorithm is used to optimal the parameter hyper-plane function; We also have analyzed the relation between software defect and attribute, and then build defect prediction model of improved particle swarm optimization to draw classify rule from sample data. Experiment results show that the adaptability and effectiveness of the proposed model.
For the connection with the interactivity of fuzzy measure and fuzzy integral, it is applied to improve the performance of PSO in this paper. In the search process, not only the direction of particle search change with all particles, and the iterations auto-modified is integration all of the inertia in the iteration progress. Thus, a novel particle swarm optimization is proposed based on fuzzy measure and fuzzy integral. The result show it is feasible for improvement optimize progress.
目前,软件缺陷预测技术的研究成为软件工程和应用中重要领域之一,该项技术的研究也越来越多。针对软件缺陷预测方法对提高软件开发质量和软件开发效率具有重要的指导意义,本文首先对计算机软件缺陷预测的基本概念、意义、研究现状进行了详细的描述,在对现有的缺陷预测技术方法进行了总结概括的基础上,分析比较了已有软件缺陷预测的优点及不足,重点研究了模糊积分理论,根据模糊积分的映射功能设计了缺陷预测模型。该模型在高维空间对软件模块数据分类并建立分类超平面完成对模块的预测。针对软件中缺陷模块和非缺陷模块的比例相差较大问题采用错误分类惩罚机制调整最小错误分类率的模糊积分分类函数。最优化分类超平面函数参数优化根据遗传优化方法来实现。其次,详细分析了软件模块属性特征与软件缺陷的关系,运用改进粒子群优化算法建立软件缺陷的预测模型。该模型通过优化提取缺陷模块特征属性规则进行预测。实验结果表明了所设计模型的有效性和适用性。
针对模糊测度和模糊积分所具有的特性,本文还定义了在粒子群优化过程中粒子个体间的交互作用和在迭代过程中惯性权重的综合运用。在迭代过程中,粒子的搜索方向及速度随着所有粒子之间关系的变化而变化,惯性权重的改变是通过综合迭代过程中的原有的惯性权重所占比重自动修改。实验结果表明基于模糊积分的粒子群优化算法是可行的。
Along with the extensive application of computers, the demand for computer software to grow. How to develop high-quality software becomes the key question that software company pay close attention to. But, the early concepts and methods of software development have limited the development of quality and development cycle of the computer in a large extent. Software defect prediction is the research that to improve software development efficiency, reduce development costs and improve the quality of software development in the field of software engineering. Study of the defect modules of software, predict erroneous software modules by classify and statistical analysis methods. The predict result is used as a reference, software developers will focus quality assurance activities on defect-prone modules and thus improve software quality and shortening development cycle by using resources more efficiently in the development process.
Now, software defect predict research is becoming an important field in software engineering and applications. And many researchers have already focused on it. Software defect prediction method has great consultive value and important significance to improving software quality and efficiency of software development. The paper begins with an elaborate of basic conceptions, significance and research status about software defect predict based on the survey of the key predict method, and analyzed the disadvantages of existing prediction methods, then researches the fuzzy integral theory, we put forward a software defect predict model based on the mapping of fuzzy integral. It adopts method with ability of clusters from data which has high dimensions to establish the classification hyper-plane in high dimensional space. And draw into the minimum ratio of misclassification to improve the classification. Simulated genetic algorithm is used to optimal the parameter hyper-plane function; We also have analyzed the relation between software defect and attribute, and then build defect prediction model of improved particle swarm optimization to draw classify rule from sample data. Experiment results show that the adaptability and effectiveness of the proposed model.
For the connection with the interactivity of fuzzy measure and fuzzy integral, it is applied to improve the performance of PSO in this paper. In the search process, not only the direction of particle search change with all particles, and the iterations auto-modified is integration all of the inertia in the iteration progress. Thus, a novel particle swarm optimization is proposed based on fuzzy measure and fuzzy integral. The result show it is feasible for improvement optimize progress.
引文
[1]于波,姜艳等,软件质量管理实践-软件缺陷预防、清除、管理实用方法,电子工业出版社,北京,2008.
[2]S Dick, A Meeks, M Last, H Bunke, A Kandel, Data mining in software metrics databases. Fuzzy Sets and Systems, Vol.145, Issue 1,2004,81-110.
[3]I Gondra, Applying machine learning to software fault-proneness prediction, Journal of Systems and Software, Vol.81, Issue 2,2008,186-195.
[4]A Erik, C Lionel, Briand, B Eivind. Johannessen, A systematic and comprehensive investigation of methods to build and evaluate fault prediction models, Journal of Systems and Software, Vol.83, Issue 1,2010,2-17.
[5]O E Karim, O EMahmoud, Predicting defect-prone software modules using support vector machines, The Journal of Systems and Software, Vol.81, Issue 5,2008, 649-660.
[6]T Khoshgoftaar, E Allen, J Deng, Using regression trees to classify fault-prone software modules, IEEE Transactions on Reliability 5, Vol.1, Issue 4,2002,455-462.
[7]C Parag, Pendharkar, Exhaustive and heuristic search approaches for learning a software defect prediction model, Engineering Applications of Artificial Intelligence Vol.23, Issue 1,2010,34-40.
[8]P C P Chang, C P Chu, Y F Yeh, Integrating in-process software defect prediction with association mining to discober defect pattern, Information and Software Technology, Vol.51, Issue 2,2009,375-384.
[9]M Cusumano, A MacCormack, C Kemerer, W Crandall, Software development worldwide:the state of practice, IEEE Computer Society,2004, Vol.20, Issue 6, 28-34.
[10]V Olivier, M David. Mining software repositories for comprehensible software fault prediction model, The Journal of Systems and Software, Vol.81, Issue 5,2008, 823-839.
[11]J S Chen, Ching-Hsue Cheng, Extracting classification rule of software diagnosis using modified MEPA, Expert Systems with Applications, Vol.34, Issue 1,2008, 411-418.
[12]J N Pizzi. Software quality prediction using fuzzy integration:a case study, Soft Computing-A Fusion of Foundations, Methodolgies and Applications, Vol.12, Issue 1,2008,67-76.
[13]B David, F Henry, L Dawn, P Maurizio, Increasing diversity:Natural language measures for software fault prediction, Journal of Systems and Software, Vol.82, Issue 11,2009,1793-1803.
[14]C Cagatay, Software Fault Prediction:A Literature Review and Current Trends, Expert Systems with Applications, Available online 30 October 2010.
[15]王青,伍书剑,李明树,软件缺陷技术,软件学报,Vol.19,No.7,2008,1565-1580.
[16]吴超,许建平,陈丽容,基于生命周期的软件缺陷预测技术,计算机工程与设计,Vol.30,No.12,2009,2956-2959.
[17]朱治雄,软件缺陷分析与预防技术研究与应用,上海交通大学硕士论文,2008
[18]JC Munson, TM Khoshgoftaar, The detection of fault-prone programs, IEEE Trans. on Software Engineering, Vol.18, Issue 5,1992,423-433.
[19]S Zhong, TM Khoshgoftaar, Seliya N. Analyzing software measurement data with clustering techniques. IEEE Intelligent Systems, Vol.19, Issue 2,2004,20-27.
[20]A Koru, H Liu, Building effective defect-prediction models in practice, IEEE Software, Vol.22, Issue 6,2005,23-29.
[21]I Witten, E Frank, Data Mining:Practical Machine Learning Tools and Techniques, second. Morgan Kaufmann, San Francisco,2005.
[22]http://mdp.iw.nasa.gov/mdp_glossary.html
[23]王熙照,模糊测度和模糊积分及在分类技术中的应用,科学出版社,北京,2008.
[24]哈明虎,模糊测度与模糊积分理论,科学出版社,北京,1998.
[25]Z.Y.Wang and G.J.Klu, Fuzzy Measure Theory, New York,1992.
[26]R Yang, Z Wang, P Heng, K Leung, Fuzzified Choquet integral with fuzzy-valued integrand and its application on temperature prediction, IEEE Transactions on Systems, Man, and Cybernetics, Vol.38, Issue 2,2008,367-380.
[27]K Xu, Z Wang, P A Heng, K S Leung, Classification by nonlinear integral projections, IEEE Trasactions on Funzzy Systems, Vol.11, Issue 2,2003,187-201.
[28]Z Wang, R Yang, P Heng, K Leung, Real-valued Choquet integrals with fuzzy-valued integrand, Fuzzy Sets and Systems, Vol.157, Issue 2,2006,256-269.
[29]A. Temko, D Macho, C Nadeu, Fuzzy integral based information fusion for classification of highly confusable non-speech sounds, Pattern Recognition, Vol.41, Issue 5,2008,1814-1823,
[30]F Hua, M L Rizzo, Honggang Wang, Kimberly Andrews Espy, Zhenyuan Wang, A new nonlinear classifier with a penalized signed fuzzy measure using effective genetic algorithm, Pattern Recognition, Vol.43, Issue 4,2010,1393-1401.
[31]M Grabisch, J M Nicolas, Classification by fuzzy integral:performance and tests, Fuzzy sets and Systems, Vol.65, Issues 2-3,1994,255-271.
[32]T Kvan, Collaborative design:what is it?, Automation in Con2straction, Vol.9, Issues 4,2000,409-415.
[33]J Robert, Kernel Entropy Componet Analysis,IEEE transactions on pattern analysis and machine intelligence, Vol.32, Issue 5,2010,847-860.
[34]J Kennedy, R C Eberhart, Particle swarm optimization, Proc. IEEE International Conference on Neural Networks IV, Australia,1995,1942-1948.
[35]J Kennedy. The Particle Swarm:Social Adaptation of Knowledge. Proceedings of IEEE International Conference on Evolutionary Computation, Indianapolis, Indiana, 1997.
[36]Y Shi, R C Eberhart, A Modified Particle Swarm Optimization, In:Proceedings of the 1999 Congress on Evolutionary Computation, IEEE Press, Piscataway, NJ,1998, 69-73.
[37]J Kennedy, R Eberhart, and Y Shi, Swarm Intelligence, Morgan Kaufman Publishers, San Francisco,2001.
[38]Y Shi, R C Eberhart, Empirical study of particle swarm optimization, Proceedings of the IEEE Congress on Evolutionary Computation, Piscataway, IEEE Press,1999, 1945-1950.
[39]张选平等.一种动态改变惯性权重的自适应粒子群算法,西安交通大学学报,Vol.39,NO.10,2005,1039-1042.
[40]王存睿,段晓东,刘向东,周福才,中值基本粒子群优化算法,辽宁省第二届学术年会暨第五届青年学术年会论文集,辽宁省第二届学术年会暨第五届青年学术年会,沈阳,2005,714-718.
[41]李烨,尹汝泼,蔡云泽等,基于离散化的支持向量机特征选择,计算机工程,Vol.32,NO.11,2006,16-17.
[2]S Dick, A Meeks, M Last, H Bunke, A Kandel, Data mining in software metrics databases. Fuzzy Sets and Systems, Vol.145, Issue 1,2004,81-110.
[3]I Gondra, Applying machine learning to software fault-proneness prediction, Journal of Systems and Software, Vol.81, Issue 2,2008,186-195.
[4]A Erik, C Lionel, Briand, B Eivind. Johannessen, A systematic and comprehensive investigation of methods to build and evaluate fault prediction models, Journal of Systems and Software, Vol.83, Issue 1,2010,2-17.
[5]O E Karim, O EMahmoud, Predicting defect-prone software modules using support vector machines, The Journal of Systems and Software, Vol.81, Issue 5,2008, 649-660.
[6]T Khoshgoftaar, E Allen, J Deng, Using regression trees to classify fault-prone software modules, IEEE Transactions on Reliability 5, Vol.1, Issue 4,2002,455-462.
[7]C Parag, Pendharkar, Exhaustive and heuristic search approaches for learning a software defect prediction model, Engineering Applications of Artificial Intelligence Vol.23, Issue 1,2010,34-40.
[8]P C P Chang, C P Chu, Y F Yeh, Integrating in-process software defect prediction with association mining to discober defect pattern, Information and Software Technology, Vol.51, Issue 2,2009,375-384.
[9]M Cusumano, A MacCormack, C Kemerer, W Crandall, Software development worldwide:the state of practice, IEEE Computer Society,2004, Vol.20, Issue 6, 28-34.
[10]V Olivier, M David. Mining software repositories for comprehensible software fault prediction model, The Journal of Systems and Software, Vol.81, Issue 5,2008, 823-839.
[11]J S Chen, Ching-Hsue Cheng, Extracting classification rule of software diagnosis using modified MEPA, Expert Systems with Applications, Vol.34, Issue 1,2008, 411-418.
[12]J N Pizzi. Software quality prediction using fuzzy integration:a case study, Soft Computing-A Fusion of Foundations, Methodolgies and Applications, Vol.12, Issue 1,2008,67-76.
[13]B David, F Henry, L Dawn, P Maurizio, Increasing diversity:Natural language measures for software fault prediction, Journal of Systems and Software, Vol.82, Issue 11,2009,1793-1803.
[14]C Cagatay, Software Fault Prediction:A Literature Review and Current Trends, Expert Systems with Applications, Available online 30 October 2010.
[15]王青,伍书剑,李明树,软件缺陷技术,软件学报,Vol.19,No.7,2008,1565-1580.
[16]吴超,许建平,陈丽容,基于生命周期的软件缺陷预测技术,计算机工程与设计,Vol.30,No.12,2009,2956-2959.
[17]朱治雄,软件缺陷分析与预防技术研究与应用,上海交通大学硕士论文,2008
[18]JC Munson, TM Khoshgoftaar, The detection of fault-prone programs, IEEE Trans. on Software Engineering, Vol.18, Issue 5,1992,423-433.
[19]S Zhong, TM Khoshgoftaar, Seliya N. Analyzing software measurement data with clustering techniques. IEEE Intelligent Systems, Vol.19, Issue 2,2004,20-27.
[20]A Koru, H Liu, Building effective defect-prediction models in practice, IEEE Software, Vol.22, Issue 6,2005,23-29.
[21]I Witten, E Frank, Data Mining:Practical Machine Learning Tools and Techniques, second. Morgan Kaufmann, San Francisco,2005.
[22]http://mdp.iw.nasa.gov/mdp_glossary.html
[23]王熙照,模糊测度和模糊积分及在分类技术中的应用,科学出版社,北京,2008.
[24]哈明虎,模糊测度与模糊积分理论,科学出版社,北京,1998.
[25]Z.Y.Wang and G.J.Klu, Fuzzy Measure Theory, New York,1992.
[26]R Yang, Z Wang, P Heng, K Leung, Fuzzified Choquet integral with fuzzy-valued integrand and its application on temperature prediction, IEEE Transactions on Systems, Man, and Cybernetics, Vol.38, Issue 2,2008,367-380.
[27]K Xu, Z Wang, P A Heng, K S Leung, Classification by nonlinear integral projections, IEEE Trasactions on Funzzy Systems, Vol.11, Issue 2,2003,187-201.
[28]Z Wang, R Yang, P Heng, K Leung, Real-valued Choquet integrals with fuzzy-valued integrand, Fuzzy Sets and Systems, Vol.157, Issue 2,2006,256-269.
[29]A. Temko, D Macho, C Nadeu, Fuzzy integral based information fusion for classification of highly confusable non-speech sounds, Pattern Recognition, Vol.41, Issue 5,2008,1814-1823,
[30]F Hua, M L Rizzo, Honggang Wang, Kimberly Andrews Espy, Zhenyuan Wang, A new nonlinear classifier with a penalized signed fuzzy measure using effective genetic algorithm, Pattern Recognition, Vol.43, Issue 4,2010,1393-1401.
[31]M Grabisch, J M Nicolas, Classification by fuzzy integral:performance and tests, Fuzzy sets and Systems, Vol.65, Issues 2-3,1994,255-271.
[32]T Kvan, Collaborative design:what is it?, Automation in Con2straction, Vol.9, Issues 4,2000,409-415.
[33]J Robert, Kernel Entropy Componet Analysis,IEEE transactions on pattern analysis and machine intelligence, Vol.32, Issue 5,2010,847-860.
[34]J Kennedy, R C Eberhart, Particle swarm optimization, Proc. IEEE International Conference on Neural Networks IV, Australia,1995,1942-1948.
[35]J Kennedy. The Particle Swarm:Social Adaptation of Knowledge. Proceedings of IEEE International Conference on Evolutionary Computation, Indianapolis, Indiana, 1997.
[36]Y Shi, R C Eberhart, A Modified Particle Swarm Optimization, In:Proceedings of the 1999 Congress on Evolutionary Computation, IEEE Press, Piscataway, NJ,1998, 69-73.
[37]J Kennedy, R Eberhart, and Y Shi, Swarm Intelligence, Morgan Kaufman Publishers, San Francisco,2001.
[38]Y Shi, R C Eberhart, Empirical study of particle swarm optimization, Proceedings of the IEEE Congress on Evolutionary Computation, Piscataway, IEEE Press,1999, 1945-1950.
[39]张选平等.一种动态改变惯性权重的自适应粒子群算法,西安交通大学学报,Vol.39,NO.10,2005,1039-1042.
[40]王存睿,段晓东,刘向东,周福才,中值基本粒子群优化算法,辽宁省第二届学术年会暨第五届青年学术年会论文集,辽宁省第二届学术年会暨第五届青年学术年会,沈阳,2005,714-718.
[41]李烨,尹汝泼,蔡云泽等,基于离散化的支持向量机特征选择,计算机工程,Vol.32,NO.11,2006,16-17.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |