GitHub Issue Bot

Overview

Implementation

Dataset

Data Preprocessing

Research Papers

Rocchio Classifier

Decision Tree Classifier

Random Forest Classifier

KNN

Naive Bayes Classifier

Bagging

Boosting

SVM

ANN with BOW

RNN with LSTM

RCNN

Summary

Project Overview

GitHub is the world’s largest platform for open source development, where developers can manage and review their code and contribute to repositories. Contributors can open issues in repositories highlighting some problem with the codebase or the documentation. An issue is created with a title, a brief description of the issue. Issues can be classified using labels such as ‘bug’, ‘feature’, ‘documentation’, etc. Such classification makes it easy for the maintainers to triage issues, for example, by giving priority to resolving a ‘bug’ over ‘documentation’. Presently, most issues have to be labelled manually, which is a laborious task. An issue labelling bot would automatically label issues when they are created, making it easier to handle issue creation and prioritization.