DEVELOPMENT OF FUZZY-BASED MODEL FOR SOFTWARE DEVELOPMENT COST ESTIMATION

Abstract

Achieving high accuracy in software development cost estimation is important in software engineering management. In this research, Fuzzy methodology is adopted to develop a software development cost estimation model which adapted fourteen software attributes of Constructive Cost Model (COCOMO) whilst identifying four more attributes that affect development effort. Data analysis was carried out on results from sampled survey from fifteen respondent software firms and experimentation on choice of fuzzy composition was done using sample projects from 93 software projects of National Aeronautical and Space Administration (NASA). A fuzzy-based software cost estimation model comprising of seventeen input variables was developed using thirty software projects from respondent organizations and was implemented in MATLAB Fuzzy Toolbox.Performance evaluation on the model based on mean magnitude relative error (MMRE) and prediction accuracy (Pred(l)) showed results of 0.17 and 82% respectively. Results from the experiments showed that the performance of a fuzzy-based software cost estimation model utilizing Takagi-Sugeno inference with Gaussian/Sigmoid membership function and imploring more number of input variables and linguistics variables is more efficient.