Software structure evolution and relation to subgraph defectiveness

Translate original post with Google Translate

Student: Ana Vranković

About project

This project was the  continuation of  ”Analysis of open source softwares using motifs” project.  The goal of this project was to study relation of subgraph frequencies in respect to subgraph defectiveness. To analyse defectiveness of the subgraphs we used statistic tests to see if there was any correlation between the two. We observed what happens with subgraph defectiveness across versions of system evolution and for different subgraph types, and if the subgraph defectiveness for different subgraph types behaves differently during system evolution.

Motivation

In this project we were motivated by this previous findings and we want to extend analysis on relation of subgraph frequences and subgraph defectiveness. We aim to observe what happens with subgraphs defectiveness across versions of system evolution and for different subgraph types. Moreover, do the subgraph defectiveness for different subgraph types behave differently across system evolution.

Results

Project has shown that software programs have similar behavior in terms of average subgraph type defectiveness and distributions of average subgraph frequencies come from the same population. However, the defectiveness of different subgraph types behaves differently and we can not find evidence that they come from the same population. During this research we found that subgraph defectiveness are good predictors of the number of defects in system version.

Research graphs results: http://ge.tt/5CgPbkY2

Comments are closed.