Method for preventing defects of electric-power information hardware

Abstract

The invention provides a method for preventing defects of electric-power information software. The method comprises the following steps of: inputting measurement information for the electric-power information at all stages in the life cycle and measurement information of the development process into an AODE prediction model, and calculating the defect densities of all stages; then searching a defect knowledge base of the electric-power information software and acquiring defect introduction reasons and corresponding prevention measures at all the stages; and finally carrying out defect prevention in a targeted manner to reduce the defect density. The method provided by the invention has the advantages that the method and the technical basis can be provided for preventing the defects in construction of an electric information system, the introduction of the defects can be prevented at the early construction of the electric information system, the defect introduction rate can be reduced, the repeated work of 'defect introduction-testing-modification' can be reduced, the targeted property for preventing the defects can be effectively improved, the project development expense of the information system can be saved, the failure risk of a software system can be reduced and the whole construction quality of the electric-power information software can be wholly improved.

Description

technical field [0001] The invention relates to a prevention method, in particular to a method for prevention of electric power information software defects. Background technique [0002] Defect prediction is the basis of prevention, and accurate prediction can focus limited resources on targeted prevention of defects. According to different prediction techniques, defect prediction methods can be divided into three categories: analogy method, Delph estimation method, and mathematical prediction model method. However, because the first two methods have insurmountable limitations, the current research work in the field of software defect prediction Part is focused on mathematical predictive modeling methods. [0003] Common mathematical prediction model methods are: Linear Discriminant Analysis (LDA), Boolean Discriminant Function (Boolean Discriminant Function, BDF), Bayesian Network (BayesianNetwork, BN), classification and regression tree (Classification And Regression Tre...

AI Technical Summary

Problems solved by technology

[0004] So far, more than 100 defect prediction models at home and abroad have been published in various professional journals and academic conferences, but most of the models have not been supported and verified by sufficient project data, and their effectiveness and accuracy have not been strengthened. ground proof, and thus not widely used

Images