A Regression Model with Mamdani Fuzzy Inference System for Early Software Effort Estimation Based on Use Case Diagrams

Effective software effort estimation is one of the biggest challenges in software engineering. One of these challenges occurs when it is required to estimate software effort in the early stages of the software life cycle, as software requirements in this stage are usually incomplete. As Unified Modeling Language (UML) model became more prominent in software requirements and design processes, software estimators became more interested to use UML model and especially the use case diagrams to estimate software. In this paper, a novel regression model based on the use case point method is created for software effort estimation. Moreover, a Mamdani Fuzzy Inference System (FIS) approach is applied on this model to enhance the accuracy of estimation. Results show that an overall improvement of 10% can be achieved after applying the proposed approach. KeywordsMamdani FIS, regression model, software effort estimation, use case diagrams, Use Case Points