Research On Measurement Technology Of Correlation Coefficient

With the development of computer science and technology,the problem of software quality and software products such as low productivity attracts more and more people's attention.In the field of software engineering,many techniques and methods are proposed and adopted to improving software quality,but the software testing are always the main way to ensure software quality.At present,defect correlation has become an important research direction of software testing,and an good technology with defect correlation can greatly reduce the testing effort and testing costs,moreover improve the efficiency of defect elimination.Therefore,the software testing strategy associated with defect correlation has a great significance.Although people have recognized the connection between software failure,but this phenomenon is also a lack of in-depth research by people.The main reason is that people has not analyzed primary causes of failure,which is analyzing the reason from the perspective of defects of software.Aiming to the above-mentioned defect,this paper considers the degree of association between the precise defects of software further.The main idea is based on software testing defect management that collects defect data in the defect management,and generate defect management of database.Then on the analysis and measurement,makes the dispersed defect data orderly.At last,the relationship between defects is analyzed to improving the quality of software that using a variety of methods for data processing This paper makes in-depth study based on the existing research on the basis of the defect correlation coefficient technology,including the following aspects:,mainly includes the following several aspects:(1)To study the method of orthogonal defect classification in defect management system and the similarity measurement method of cluster analysis.Aiming at the shortcomings of the existing defects of correlation measurement method in defect metrics for transmitting information,build a kind of defect data similarity model based on the measurement of orthogonal defect classification method and similarity clustering analysis method.The model is similar to the actual software testing process,and having a higher practicability.(2)This paper studies the relationship between the defect management system deficiencies a nd the associated components.Through research and analysis of defects in the process and the relationship between the object-oriented component transmission,designed a kind of coupling between components strategy,to indirectly measure the coupling between defects.(3)Study the spread of defects and defect correlation coefficient transmission strategy.By establishing the relationship among the similarity of data defects,defective coupling of software defects and the relationship of the correlation coefficient of software defects,establish a defect association rules based on defective deliver.Finally,a strategy based on correlation coefficient of defect transfer is designed to achieve adjusting correlation coefficient and updating associated matrix of defects.Finally,a series of simulation experiments are made to verify the similarity between data defects and defective coupling between the defect correlation coefficients measure are contribution to the process of defect correlation coefficient.Experimental results show that the proposed method could complete the same quality defects in a shorter period of time in the condition of the task of equal quality and improve testing efficiency compared with the traditional method and defect correlation coefficient associated defect management methods and strategies.