Loading...

Analyzing Open Source Software Development Networks Using Community Detection Methods

Hamedi, Mohammad Hossein | 2015

632 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: English
  3. Document No: 47290 (52)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Khansari, Mohammad
  7. Abstract:
  8. Free and Open-Source Software (FOSS) is grown rapidly and many researches are conducted to study FOSS developments. The studies aim to help FOSS development by finding bugs, analyzing developers, modeling evolution of software and categorizing them. Finding proper developers or key developers of each group is essential in developing software, recommend developers and assigning bugs to proper developers. Community detection detects groups of nodes having dense connections. The groups represent developers which cooperate closely on developing FOSS. The thesis proposes a method based on evolutionary algorithms and topological analysis to detect communities in open-source software development networks. Evolutionary models search to find optimal solutions and commonly have better results than greedy algorithms. But they have more time complexity. In our problem number of nodes is not too high and we expect to have more accurate results. Genetic algorithm was applied to find communities and a new chromosome model was also proposed and applied in the genetic algorithm. The results had better modularity and runtime compared with other evolutionary models. By analyzing detected communities from an open-source software development network, the key developers were categorized based on their impact on the criteria and their links to the other authors and projects. Their concentrations have been studied along with the main focus of each detected community
  9. Keywords:
  10. Evolutionary Algorithm ; Community Detection ; Social Networks ; Free And Opensource Software ; Opensource Development Network

 Digital Object List

 Bookmark

No TOC