Evaluate interoperability of foundational software platform by Bayesian networks
-
摘要: 互操作性是当前软件最重要的特性之一.通过分析问题域,结合贝叶斯网的特征域,提出使用贝叶斯网来解决基础软件平台的互操作性评估问题.首先根据问题域选取贝叶斯算法,并收集实际数据以引入与问题相关的领域知识.根据所选取的算法构造互操作性的贝叶斯网结构,并且进一步学习此结构的参数.在此过程中,对选取的K2算法进行改进.然后,利用贝叶斯推理来根据互操作性的结构和参数得出评估对象的互操作性等级.最后,一个实例讲述了方法具体的应用过程.实验结果证明了方法的合理性.Abstract: Interoperability is one of the most important features of current software. Bayesian networks was proposed to solve the problem of evaluating interoperability of foundational software platform after analyzing both the problem domain and the Bayesian networks feature domain. First of all, Bayesian algorithm was chosen according to the problem domain, and actual data were gathered for importing domain knowledge. Bayesian structure was formed by the selected algorithm and used to learn its parameters. During this process, the chosen K2 algorithm was improved. After that, Bayesian reasoning was used to evaluate the grade of interoperability of foundational software platform according to the established Bayesian structure and parameters. At last, an application showed how to use this brand-new model to evaluate the interoperability of foundational software platform. The experimental result proved the reasonableness and usefulness of the model.
-
Key words:
- Bayesian network /
- interoperability /
- foundational software platform /
- K2 algorithm
-
[1] Lu Shouqun.Discussion on interoperability of open-source software [J]. Software World,2007, 32(6):1-50 [2] Komatsoulis G A. Introduction to interoperability and compatibility . National Cancer Institute Centre for Bioinformatics, 2006. https://cabig.nci.nih.gov/events/2006/2006_AnnualMeeting_Day_2/Intro_Interoperability_Compatibility_Komatsoulis.pdf [3] PIT Ltd. PIT Philosophy .2007 .http://www.qpit.ltd.uk/media.papers.Defining-Quality.pdf [4] 高晖.软件体系结构层次的软件适应性度量和预测研究 .北京:北京航空航天大学计算机学院,2008 Gao Hui. Research on flexibility metrics in software architecture level . Beijing: School of Computer Science and Technology, Beijing University of Aeronautics and Astronautics,2008 [5] Herskovits E. Computer-based probabilistic-network constructi-on . CA: Stanford University, 1991 [6] Dagum P, Luby M. Approximating probabilistic reasoning in Bayesian belief networks is NP-hard [J]. Artificial Intelligence, 1993, 60(1): 51-91 [7] Mitchell T M. Machine learning [M]. Beijing: China Machine Press, 2003: 86-154 [8] Guo H, Hsu H. A survey of algorithms for real-time Bayesian network inference . 2002. http://citeseer.ist.psu.edu/guosurvey.html [9] Fan C F,Yu Y C.BBN-based software project risk management [J]. Journal of Systems Software,2004,73:193-203 [10] Stamelos I,Angelis L,Dimou P,et al. On the use of Bayesian belief networks for the prediction of software productivity[J]. Information and SoftwareTechniques,2003,45:51-60 [11] Fenton N E,Neil M,Marsh W,et al. Predicting software defects in varying development lifecycles using Bayesian nets [J]. Information and Software Technology,2007,49:32-43 [12] Cooper G, Herskovits E. A Bayesian method for the induction of probabilistic networks from data [J]. Machine Learning,1992,9:309-347 [13] Microsoft Research. MSBNx . 2006.http://research.microsoft. com/adapt/MSBNx [14] KSU Probabilistic Reasoning Group. BNJ . 2004.http://bnj. sourceforge.net -
点击查看大图
计量
- 文章访问数: 3544
- HTML全文浏览量: 101
- PDF下载量: 1310
- 被引次数: 0
下载:
