Abstract:
When developers use the code completion in their Integrated Development Environment (IDE), they provide implicit feedback about the usage of the Application Programming Interfaces (APIs) they program against. We demonstrate how to apply Collaborative Filtering techniques to compute context-sensitive completion recommendations from such feedback and discuss how the approach can be used to bring the knowledge of the crowd to every developer.
Resources
BibTeX
@inproceedings {APM14,
title = {{Method-Call Recommendations from Implicit Developer Feedback}},
author = {Amann, Sven and Proksch, Sebastian and Mezini, Mira},
booktitle = {{Proceedings of the 1st International Workshop on CrowdSourcing in Software Engineering}},
series = {CSI-SE 2014},
pages = {5--6},
year = {2014},
doi = {10.1145/2593728.2593730},
url = {http://dx.doi.org/10.1145/2593728.2593730},
}