Multi-instance learning for software quality estimation in object-oriented systems: a case study

2010 ◽  
Vol 11 (2) ◽  
pp. 130-138 ◽  
Author(s):  
Peng Huang ◽  
Jie Zhu
Author(s):  
Yann-Gaël Gueheneuc ◽  
Jean-Yves Guyomarc’h ◽  
Khashayar Khosravi ◽  
Hourari Sahraoui

Software quality models link internal attributes of programs with external quality characteristics. They help in understanding relationships among internal attributes and between internal attributes and quality characteristics. Object-oriented software quality models usually use metrics on classes (such as number of methods) or on relationships between classes (for example coupling) to measure internal attributes of programs. However, the quality of object-oriented programs does not depend on classes solely: it depends on the organisation of classes also. We propose an approach to build quality models using patterns to consider program architectures. We justify the use of patterns to build quality models, describe the advantages and limitations of such an approach, and introduce a first case study in building and in applying a quality model using design patterns on the JHotDraw, JUnit, and Lexi programs. We conclude on the advantages of using patterns to build software quality models and on the difficulty of doing so.


2021 ◽  
Vol 12 (3) ◽  
pp. 1-16
Author(s):  
Mokhtaria Bouslama ◽  
Mustapha Kamel Abdi

The cost of software maintenance is always increasing. The companies are often confronted to failures and software errors. The quality of software to use is so required. In this paper, the authors propose a new formal approach for assessing the quality of object-oriented system design according to the quality assessment model. This approach consists in modeling the input software system by an automaton based on object-oriented design metrics and their relationship with the quality attributes. The model exhibits the importance of metrics through their links with the attributes of software quality. In addition, it is very practical and flexible for all changes. It allows the quality estimation and its validation. For the verification of proposed probabilistic model (automaton), they use the model-checking and the prism tool. The model-checking is very interesting for the evaluation and validation of the probabilistic automaton. They use it to approve the software quality of the three experimental projects. The obtained results are very interesting and of great importance.


2020 ◽  
Vol 5 (17) ◽  
pp. 1-5
Author(s):  
Jitendrea Kumar Saha ◽  
Kailash Patidar ◽  
Rishi Kushwah ◽  
Gaurav Saxena

Software quality estimation is an important aspect as it eliminates design and code defects. Object- oriented quality metrics prediction can help in the estimation of software quality of any defects and the chances of errors. In this paper a survey and the case analytics have been presented for the object-oriented quality prediction. It shows the analytical and experimental aspects of previous methodologies. This survey also elaborates different object-oriented parameters which is useful for the same problem. It also elaborates the problem aspects as well the limitations for the future directions. Machine learning and artificial intelligence methods have been considered mostly for this survey. The parameters considered are inheritance, dynamic behavior, encapsulation, objects etc.


Sign in / Sign up

Export Citation Format

Share Document