الفهرس | Only 14 pages are availabe for public view |
Abstract Effort estimation in the domain of software development is a process of forecasting the amount of effort expressed in persons/month required to develop software. Most of the existing effort estimation techniques are suitable for traditional software projects. The agile methodologies adopted on an idea of iterative and evolutionary development processes. The nature of agile software projects is different from traditional software projects. Moreover, the traditional estimation techniques need a well-defined requirements where the requirements in agile methodologies are subject to change; therefore using the traditional effort estimation techniques can produce inaccurate estimation. Agile software projects require an innovated effort estimation framework to help in producing accurate estimation. The goal of this thesis is the utilization of fuzzy logic in improving the effort estimation accuracy using the user stories by characterizing inputs parameters using trapezoidal membership functions. The framework is formulated to idealize the characteristics and properties of agile software development methodologies. It is based on COCOMO II factors, story point, priority factors, and velocity factors. The results evaluated via Magnitude of Relative Error (MRE) and Prediction Level (PRED) metrics. The results show the proposed framework increasing the accuracy in term of Prediction Level (PRED) |