The Analysis of Software Complexity Using Stochastic Metric Selection

Abstract

The automated prediction of qualitative attributes such as software complexity is a desirable software engineering goal. A potential technique is to use software metrics as quantitative predictors for these kinds of attributes. We describe a pattern classification method where a large collection of classifiers is presented with randomly selected subsets of software metrics describing modules from a sophisticated biomedical data analysis system. The method identifies the software metric subset that has the highest discriminatory power vis-à-vis software complexity. That is, we identify the metric subset that is most effective at predicting this qualitative attribute. This classification method is empirically evaluated and carefully validated against three benchmark approaches. We demonstrate that this method has utility in the automated prediction of software complexity using quantitative measures.

A Procedure for Extending Input Selection Algorithms to Low Quality Data in Modelling Problems with Application to the Automatic Grading of Uploaded Assignments

“…Because of this, some researchers propose to use feature selection techniques for finding the best software metrics for the problem at hand. In [ 25 ], a stochastic procedure is employed to select the subset of quantitative measures that bring out the best software quality prediction. Another example is [ 26 ], where eighteen filter-based feature selection procedures are tested against sixteen software datasets, in this case searching for fault prone modules.…”

Incremental Feature Selection Method for Software Defect Prediction

“…They used COCOCMO-1, COCOCMO-2 datasets and recommended the feature selection in cost modeling, particularly when dealing with very small datasets and concluded that reduced datasets could improve the performance. Pizzi et al [17] explained a stochastic metric selection method identifies subset which is most effective in prediction of software module complexity. Three benchmark datasets have been used for evaluation of classification method.…”