Download PDFOpen PDF in browser

Open Source Software Survivability Prediction Using Multi Layer Perceptron

10 pagesPublished: September 26, 2019

Abstract

Many organizations develop software systems using Open Source Software (OSS) components. OSS components have a high risk of going out of support, making dependency on OSS components risky. So, it is imperative to perform risk assessment during the selection of OSS components. A model that can predict OSS survivability provides an objective criterion for such an assessment. Currently, there are no simple, quick and easy methods to predict survivability of OSS components. In this paper, we build a simple Multi Layer Perceptron (MLP) to predict OSS survivability. We performed experiments on 449 OSS components containing 215 active components and 234 inactive components collected from GitHub. The evaluation results show MLP achieves 81.44% validation accuracy for survivability prediction on GithHub dataset.

Keyphrases: github, machine learning, mlp, open source software, software complexity

In: Frederick Harris, Sergiu Dascalu, Sharad Sharma and Rui Wu (editors). Proceedings of 28th International Conference on Software Engineering and Data Engineering, vol 64, pages 148-157.

BibTeX entry
@inproceedings{SEDE2019:Open_Source_Software_Survivability,
  author    = {Vijaya Kumar Eluri and Shahram Sarkani and Thomas Mazzuchi},
  title     = {Open Source Software Survivability Prediction Using Multi Layer Perceptron},
  booktitle = {Proceedings of 28th International Conference on Software Engineering and Data Engineering},
  editor    = {Frederick Harris and Sergiu Dascalu and Sharad Sharma and Rui Wu},
  series    = {EPiC Series in Computing},
  volume    = {64},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2398-7340},
  url       = {/publications/paper/dxJl},
  doi       = {10.29007/cmc6},
  pages     = {148-157},
  year      = {2019}}
Download PDFOpen PDF in browser