Software development effort estimation using fuzzy logic: a case study

Software estimation has been identified as one of the three great challenges for half-century-old computer science. Developers should be able to achieve practices containing effort estimation based on their own programs. New paradigms as fuzzy logic may offer an alternative for software effort estimation. This paper describes an application whose results are compared with those of a multiple regression. A subset of 41 modules developed from ten programs are used as data. Result shows that the value of MMRE (an aggregation of magnitude of relative error, MRE) applying fuzzy logic was slightly higher than MMRE applying multiple regression; while the value of Pred(20) applying fuzzy logic was slightly higher than Pred(20) applying multiple regression. Moreover, six of 41 MRE was equal to zero (without any deviation) when fuzzy logic was applied (not any similar case was presented when multiple regression was applied).