Analyzing the impact of information retrieval and relevance feedback techniques on concept location tasks

Abstract: Concept location is the process by which a programmer determines the place, within a system codebase, where a change is to start in response to a modification request. It is a usual and fundamental process performed as part of software maintenance tasks such as bug fixing, refactoring, and in some cases, new feature implementation. One of the recent approaches proposed to support that process augments information retrieval (IR) based concept location via an explicit relevance feedback (RF) mechanism. In this thesis, we present an Eclipse plugin that implements the IR+RF approach and a controlled experiment aimed at assessing the impact of that approach on bug fixing tasks. Within the experiment, five bug fixing tasks were performed by 40 undergraduate software engineering students. The efficiency of the participants, the completion time, and the correctness of their responses were measured. The results indicate that the IR+RF approach surpasses in effectiveness and efficiency the default searching functionalities provided by the Eclipse IDE. On the other hand, it does not reduce the completion time in bug fixing tasks.

• Autores: Wilches Riaño, Andrés Fernando
• Fecha de publicación: 2017
• Institución: Universidad Nacional de Colombia
• Repositorio: Universidad Nacional de Colombia
• Idioma: spa
• OAI Identifier: oai:repositorio.unal.edu.co:unal/62402
• Acceso en línea: https://repositorio.unal.edu.co/handle/unal/62402; http://bdigital.unal.edu.co/61512/
• Palabra clave: 02 Bibliotecología y ciencias de la información / Library and information sciences; 62 Ingeniería y operaciones afines / Engineering; concept location; information retrieval; relebance feedback; controlled experiment; localización de conceptos; recuperación de la información; retroalimentación relevante; experimento controlado
• Rights: openAccess
• License: Atribución-NoComercial 4.0 Internacional