원문정보
초록
영어
The research, development and applications of software measurement have been carried out for more than forty years. Many researchers have done much in it, obtained lots of theoretical results, and developed a series of practical applications. At present, with the rapid development of object-oriented technology used in software theories and application, how to measure software in an effectively and scientifically way is becoming a hot and difficult issue. Class diagram as the most important UML model, its reasonable design has a remarkable effect on the ultimate system. This paper applies information entropy-based multi-attribute decision in UML class diagram metrics, providing a method to measure UML class diagram complexity weights. Its attributes out of information entropy and attributes’ weigh vectors can measure the complexity of object-oriented software effectively and scientifically. Besides, case analysis can prove this metric’s usability. The more precise experiment outcome has proved that the method is connected to human’s experience, and also can be applied to improve software quality.
목차
1. Introduction
2. A method of measuring UML Class Diagram Complexity Weights
2.1. Measuring Complexity of UML Class Diagrams
2.2. Case Study
2.3. Comparative Analysis of The Experimental Results
3. Conclusion
Acknowledgements
References