A tool for discriminant analysis and classification of software metrics

The potential of software complexity measurements has not been fully explored for lack of a systematic approach to collect, evaluate and use these measurements. The goal of the discriminant analysis tool and classification model for software complexity metrics is to automate the process of collecting software measurements. It also develops a methodology to determine the robustness of several combined metrics and offers a norm against which measurements can be compared. The model acts as an evaluator of the many metrics that have been and may be suggested in the literature and offers programmers a systematic approach to choose the ones that best fit their particular type of system, problem domain and environment.