This work aims at predicting the number of defects of Object Oriented (OO) software using Computational Intelligence techniques. There are 6 software metrics, also known as "CK metrics" to use as inputs for the prediction system, and the number of modifications made to the software projects as their output values. The CK metrics and number of Lines of Code of 5 software projects are available for this work, they are to be used to generate a system capable of determining the number of modifications made in the software projects based on their CK metrics.; The techniques to use in this work are: Fuzzy Clustering, Multivariable regression, Clustering and Local Regression, Neural Networks, Switching Regression Models and Fuzzy Clustering, and Genetic Algorithm - Based Clustering Method. At the end the different method are compared and discussed.
展开▼
