Clustering Support for Static Concept Location in Source Code

Authors:
Giuseppe Scanniello;Andrian Marcus
Affiliations:
-;-
Venue:
ICPC '11 Proceedings of the 2011 IEEE 19th International Conference on Program Comprehension
Year:
2011

Citing 0
Cited 3

The effects of identifier retention and stop word removal on a latent Dirichlet allocation based feature location technique

Proceedings of the 50th Annual Southeast Regional Conference
Analysis of user comments: an approach for software requirements evolution

Proceedings of the 2013 International Conference on Software Engineering
Leveraging machine learning and information retrieval techniques in software evolution tasks: summary of the first MALIR-SE workshop, at ASE 2013

ACM SIGSOFT Software Engineering Notes

Quantified Score

Hi-index	0.00

Visualization

Abstract

One of the most common comprehension activities undertaken by developers is concept location in source code. In the context of software change, concept location means finding locations in source code where changes are to be made in response to a modification request. Static techniques for concept location usually rely on searching the source code using textual information or on navigating the dependencies among software elements. In this paper we propose a novel static concept location technique, which leverages both the textual information present in the code and the structural dependencies between source code elements. The technique employs a textual search in that source code, which is clustered using the Border Flow algorithm, based on combining both structural and textual data. We evaluated the technique against a text search based baseline approach using data on almost 200 changes from five software systems. The results indicate that the new approach outperforms the baseline and that improvements are still possible.