Recommender Systems for Software Project Managers

TL;DR

This paper reviews the development and application of recommender systems in software project management, analyzing four open source systems and implementing a customized engine to improve stakeholder coordination.

Contribution

It introduces a customized recommender engine for software engineering, evaluating existing systems and identifying key issues to enhance project stakeholder coordination.

Findings

Analysis of four open source recommender systems

Implementation of a customized recommender engine

Identification of main issues in current systems

Abstract

The design of recommendation systems is based on complex information processing and big data interaction. This personalized view has evolved into a hot area in the past decade, where applications might have been proved to help for solving problem in the software development field. Therefore, with the evolvement of Recommendation System in Software Engineering (RSSE), the coordination of software projects with their stakeholders is improving. This experiment examines four open source recommender systems and implemented a customized recommender engine with two industrial-oriented packages: Lenskit and Mahout. Each of the main functions was examined and issues were identified during the experiment.