Productivity prediction model based on Bayesian analysis and productivity console
MetadataShow full item record
Software project management is one of the most critical activities in modern software development projects. Without realistic and objective management, the software development process cannot be managed in an effective way. There are three general problems in project management: effort estimation is not accurate, actual status is difficult to understand, and projects are often geographically dispersed. Estimating software development effort is one of the most challenging problems in project management. Various attempts have been made to solve the problem; so far, however, it remains a complex problem. The error rate of a renowned effort estimation model can be higher than 30% of the actual productivity. Therefore, inaccurate estimation results in poor planning and defies effective control of time and budgets in project management. In this research, we have built a productivity prediction model which uses productivity data from an ongoing project to reevaluate the initial productivity estimate and provides managers a better productivity estimate for project management. The actual status of the software project is not easy to understand due to problems inherent in software project attributes. The project attributes are dispersed across the various CASE (Computer-Aided Software Engineering) tools and are difficult to measure because they are not hard material like building blocks. In this research, we have created a productivity console which incorporates an expert system to measure project attributes objectively and provides graphical charts to visualize project status. The productivity console uses project attributes gathered in KB (Knowledge Base) of PAMPA II (Project Attributes Monitoring and Prediction Associate) that works with CASE tools and collects project attributes from the databases of the tools. The productivity console and PAMPA II work on a network, so geographically dispersed projects can be managed via the Internet without difficulty.
Yun, Seok Jun (2003). Productivity prediction model based on Bayesian analysis and productivity console. Doctoral dissertation, Texas A&M University. Texas A&M University. Available electronically from