AI · Supervised machine-learning comparison with random forest and 10-fold cross-validation.

Predicting long-time contributors for GitHub projects using machine learning

Peer-reviewed research · 2021

Vijaya Kumar Eluri, Thomas A. Mazzuchi, Shahram Sarkani

Summary

Random forest performed best; repository and contributor metadata can predict long-time-contributor status, and the paper identifies important features for prediction.

For your marketing

Possible founder application

Use early signals to triage mentorship and follow-up, not as an automated gatekeeper for contributors.

Reported result

What the study found

Random forest performed best; repository and contributor metadata can predict long-time-contributor status, and the paper identifies important features for prediction.

Research question

What the authors investigated

Can repository/contributor features predict whether new contributors become long-time contributors, and which features matter?

Design

Supervised machine-learning comparison with random forest and 10-fold cross-validation.

Sample and setting

Contributors from 888 GitHub projects across multiple languages and application domains.

Scope and transfer

Limitations to keep in view

The authors state the statistical approach cannot establish causality; GitHub-project selection limits generalizability despite broad coverage.

Research sources

Publisher/authoritative landing page and abstract/full text fetched: https://www.sciencedirect.com/science/article/pii/S0950584921000902 Firecrawl research job 01a14366-d263-71d0-bc44-b95482987b7b returned this citation and metadata.

Study strength

3.5/5 Abstract-level evidence is useful for framing questions; verify full methods before transfer.

How this was assessed

The 1–5 evidence-check score is a practical reading aid, not a formal peer review or a probability that a conclusion is true. It combines design strength (2 points), sample reporting (1), comparison/control clarity (1), and how much of the paper could be checked (1). Access affects how confidently this directory can assess a study, not the study's underlying quality. Blinding is reported separately because it is not relevant to every design.

Design transparency · 2/2

Based on the reported study design in the retrieved source.

Sample and setting · 1/1

The sample context is stated; a numerical count is preserved only when reported.

Outcome clarity · 1/1

The finding is summarized from publisher or article metadata.

Access for checking · 0.5/1

Only abstract or bibliographic access was confirmed.

Double-blind? The retrieved publisher record does not establish whether participants or outcome assessors were blinded.