Mineração de dados aplicada ao controle de prazos e prioridades em projetos de software

Abstract

Currently, software products are produced at an industrial level and with increasing complexity, so that their production requires a considerable number of people working simultaneously and following the same processes. Among the main problems encountered in these scenarios are the proper allocation of priorities, deadlines and how to define the importance of the requirements, as well as estimate the maximum deadline for delivery so that it is produced efficiently and safely. Although there are currently project management systems, they leave only the project manager responsible for the actions to be taken in order for the project to be successfully completed. The features offered in these systems do not assist the manager in his strategic decisions, such as negotiating a longer deadline for delivery of a requirement that will not be completed in time. On the other hand, there are computational algorithms and techniques capable of assisting decision making in different scenarios. It is assumed that the use of Data Mining, combined with methodologies for management of schedules and risks, can offer managers suggestions and / or alternatives for problems encountered in the software development process. In this line, this project aimed to apply Data Mining algorithms for Knowledge Discovery in Databases, specifically Decision Tree Classification algorithms, in the development of a project management system, which will recommend management actions in management Project deadlines and priorities. Results show that the system assists managers in the designation of priorities and deadlines, through records of the manager's own behavior in critical situations, in order to obtain greater success in their projects.