FunkR-pDAE: Personalized Project Recommendation Using Deep Learning

Abstract

Open source communities are the largest
coding platforms that offer software developers
an opportunity to engage in social activities
relating to the software development. The
developers can store and share their codes or
projects with wider community of people using
the repositories. Enabling its users to find relevant
projects is the prime feature of the open source
communities. Finding a relevant project among
vast open source projects is a difficult task.
Recommending a suitable project for developers
can save their time and energy. Pengcheng Zhang
et.al. Used FunkR-pDAE, a personalized project
recommendation approach based on Deep Auto-
Encoders, deep learning model. This approach has
clear limitations, such as ignoring the dataset
description and source code comment documents.
So, proposing a method for
recommending a project in open source
communities that takes into account the dataset
description and source code comment documents,
allowing for the discovery of more relationships
between developers and projects this might in turn
lead to increased precision rate. And also in the
proposed work it is planned to generalize the
model to recommend other open source
communities i.e., github, kaggle and source forge
along with integrating the dataset description and
source code comment documents in order to mine
more relations among the developers and
projects.