Font Size: a A A

The Design And Implementation Of Software Process Risk Prediction Subsystem Of DevOps Management Platform

Posted on:2021-01-10Degree:MasterType:Thesis
Country:ChinaCandidate:K X LuFull Text:PDF
GTID:2428330647450852Subject:Software engineering
Abstract/Summary:PDF Full Text Request
There are many risk factors in the process of software development,which slow down the process of software development and even affect product quality.As the starting point and fundamental purpose of software development,requirements are very important for software development.However,during the development process,the requirement will inevitably change.Once requirements change,a lot of manpower and material resources will be wasted,which will affect the completion of the whole project.In addition,the timely delivery of products is very important for each team,which requires the team to accurately estimate the progress in the plan,identify the risks in the development,and reasonably arrange and adjust the development plan.It's a difficult problem for decision makers.However,in practice,the prediction of risk mostly depends on the personal experience of decision-makers,lacking theoretical support.In the process of Dev Ops reform,a large domestic enterprise encountered the same problems,such as lack of risk assessment in the process of software development,process control is nonstandard,etc.Therefore,a subsystem that can predict the risks in software process is particularly important,so that the project manager can dynamically adjust and manage the software development process.Base on this background,the software process risk prediction subsystem is implemented to meet the above requirements.The system uses the Spring Boot framework and provides a user-friendly front-end.The system is divided into three modules:data processing module,change risk prediction of requirement module and delivery time prediction of requirment module.Data processing module realizes data cleaning and data processing by means of missing value processing,abnormal value detection,feature standardization,etc.The change risk prediction of requirement module uses machine learning methods such as Support Vector Machine,Decision Tree,Random Forest and Naive Bayes to build models and train requirement data to realize the prediction of requirement change risk.The delivery time prediction of requirment module realizes the prediction of requirment delivery risk by modeling the software process and computing the System Dynamics engine.The system realizes the identification of the risk factors in the software process through the change risk prediction of requirement and delivery time prediction of requirment,helps the project manager to make scientific decisions,reasonably arrange and adjust the work plan,and ensures the on-time delivery of the software products.Since the system launched,the system has been running steadily and user feedback is good.Effectively help project managers anticipate risk factors,make scientific decisions,reasonably formulate work plans,adjust work arrangements,ensure timely delivery of the project,and ensure the credibility of the process.
Keywords/Search Tags:Requirement Change Prediction, Requirement Delivery Prediction, Machine Learning, Software Process Simulation
PDF Full Text Request
Related items