Fuzzy Logic Systems for Software Development Effort Estimation Based Upon Clustering of Programs Segmented by Personal Practices

The most common application of software metrics is to develop models that predict the effort that will be required to complete certain stages of a software development. There are two main stages for using an estimation model (1) it must be determined whether the model is adequate to describe the observed (actual) data, that is, the model adequacy checking; if it resulted adequate then (2) the estimation model is validated in its environment using new data. In this paper, an investigation aimed to compare personal fuzzy logic systems (FLS) with linear regression is presented. The evaluation criteria is based upon ANOVA of MRE and MER, as well as MURE, MMER and pred(25). Ninety-nine programs were developed by eighteen programmers. From these programs four FLS are generated for estimating the effort of sixty programs developed by a group of ten developers. Results show that a FLS can be used as an alternative for estimating the development effort at personal level