Development of an early software project risk asssessment application using case-based reasoning / Ahmad Zaki Kamaruddin

One of the major problems with early software project development is the difficulty to assess and manage the software project risk that might occur. It is vital for project manager to assess the risk because failure to recognize risk can cause the project to fail. Many applications have been developed in software development to assist project manager to estimate quality and cost based on relevant features of a project using the neural network and fuzzy set theory. It helps project manager to reduce the risk of incorrect estimation of project cost in software development. However, this project is focuses on the development of early software development project risk assessment using case-based reasoning (CBR) technique. It covers the six software project risk feature that are selected as case feature and the Nearest Neighbor algorithm that is used for similarity matching process in retrieval phase of the CBR cycle. The prototype has been successfully tested using new data. It shows that the CBR technique is able to find the similarity between the new input case and previous experience cases and give a recommended solution.