Abstract-Software project management is the art and
science of planning and leading software projects to achieve
predetermined corporate goals. It requires knowledge of the
entire software development lifecycle. The project manager's
main responsibility is to ensure a successful project outcome.
Project success is normally defined as achieving desired project
objectives and features within desired cost and schedule. Many
factors affect project success including dealing with gathering
requirements, customer involvements and project management.
Several researchers have investigated the success or failure of
software projects using statistical approaches.
In this paper, a web based survey and interviews are used to
collect project data, about requirements, project sponsor and
customers. Tools such as association, neural networks,
clustering, naive bayes and decision tree are used to discover
common characteristics and rules that govern project success
and failure.
The results show the power of data mining algorithms to
discover the most important factors and associations in project
success and failure. Results showed that each mining algorithm
has a particular strength to provide knowledge and make
predictions about project success opportunities.