|
|
Google Summer of Code 2011 |
|
[UPDATE:] We have finished integrating Apache Mahout into WebDSL! It is now possible to generate web applications that can produce recommendations, such as community recommendations in the case of this project. More updates soon! For this year's Google Summer of Code, I proposed an alternate project based on research in model-driven software development and recommendation systems for software engineering. I also proposed to give a talk about the project at ApacheCon NA in Vancouver. Even though my proposed project was not accepted for GSoC this year, I am still looking forward to working on this idea and implementing it in an Apache Lab. Later I would like to present the finished Apache Lab at ApacheCon NA. My proposal can be found in the link below: |
Design, implement and test a domain-specific sub-language (DSL) for collaborative filtering recommender systems in WebDSL. Propose an Apache Lab called Community recommendations that brings together the data mining capabilities of SVNPlot and Apache Mahout, in a collaborative filtering web application generated by WebDSL that issues recommendations to ASF committers about source code files they might be interested to co-author with other developers, within and/or across ASF (incubator) communities. Present community recommendation Apache Lab at ApacheCon NA 2011 in Vancouver.
1. Design, implementation and testing of a domain-specific sub-language (DSL) for collaborative filtering recommender systems in WebDSL.
2. Application of the WebDSL sub-language in the generation of a collaborative filtering web application Community recommendations on Apache Labs.
3. Data collection for research into the effects of recommendations on code production in open source communities and comparison to a system under development for recommendations on content production in Twitter communities.