Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura

Background: The decisions made by architects to favor short-term objectives and potentially to the detriment of long-term software quality are known as Technical Debt in Architecture. This type of technical debt is difficult to identify because it is related to quality attributes that are not visibl...

Full description

Saved in:
Bibliographic Details
Main Authors: Boris Rainiero Pérez-Gutiérrez, Darío Ernesto Correal, Fredy Humberto Vera-Rivera
Format: Article
Language:English
Published: Fundación de Estudios Superiores Comfanorte 2022-01-01
Series:Mundo Fesc
Subjects:
Online Access:https://www.fesc.edu.co/Revistas/OJS/index.php/mundofesc/article/view/1198
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850031721445064704
author Boris Rainiero Pérez-Gutiérrez
Darío Ernesto Correal
Fredy Humberto Vera-Rivera
author_facet Boris Rainiero Pérez-Gutiérrez
Darío Ernesto Correal
Fredy Humberto Vera-Rivera
author_sort Boris Rainiero Pérez-Gutiérrez
collection DOAJ
description Background: The decisions made by architects to favor short-term objectives and potentially to the detriment of long-term software quality are known as Technical Debt in Architecture. This type of technical debt is difficult to identify because it is related to quality attributes that are not visible to the client, such as maintainability and the system's ability to evolve. Objective: Therefore, in this article, a supervised machine learning model is used to support the identification of technical debt in architecture, located at the architectural design stage. Methods: This proposal relies on information collected from artifacts produced during the architectural design process to build a dataset that allows the evaluation of different supervised learning algorithms and thus establish the one with the best accuracy. The identification of technical debt, within the framework of this proposal and unlike those proposed in the literature, does not consider the source code. Results: The performance of the model was evaluated through a real-life industrial case study, revealing that both accuracy and recall were at acceptable levels. Conclusions: The data used to train the model, while appropriate, is susceptible to further improvement. This approach will allow architects to support the identification of conscious and unconscious ATDs injected into their architectures.
format Article
id doaj-art-ce94bc1dea9c47e7acc689eb785833c6
institution DOAJ
issn 2216-0353
2216-0388
language English
publishDate 2022-01-01
publisher Fundación de Estudios Superiores Comfanorte
record_format Article
series Mundo Fesc
spelling doaj-art-ce94bc1dea9c47e7acc689eb785833c62025-08-20T02:58:54ZengFundación de Estudios Superiores ComfanorteMundo Fesc2216-03532216-03882022-01-01122314415710.61799/2216-0388.1198Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitecturaBoris Rainiero Pérez-Gutiérrez0Darío Ernesto Correal1Fredy Humberto Vera-Rivera2Universidad Francisco de Paula SantanderUniversidad de los AndesUniversidad Francisco de Paula SantanderBackground: The decisions made by architects to favor short-term objectives and potentially to the detriment of long-term software quality are known as Technical Debt in Architecture. This type of technical debt is difficult to identify because it is related to quality attributes that are not visible to the client, such as maintainability and the system's ability to evolve. Objective: Therefore, in this article, a supervised machine learning model is used to support the identification of technical debt in architecture, located at the architectural design stage. Methods: This proposal relies on information collected from artifacts produced during the architectural design process to build a dataset that allows the evaluation of different supervised learning algorithms and thus establish the one with the best accuracy. The identification of technical debt, within the framework of this proposal and unlike those proposed in the literature, does not consider the source code. Results: The performance of the model was evaluated through a real-life industrial case study, revealing that both accuracy and recall were at acceptable levels. Conclusions: The data used to train the model, while appropriate, is susceptible to further improvement. This approach will allow architects to support the identification of conscious and unconscious ATDs injected into their architectures.https://www.fesc.edu.co/Revistas/OJS/index.php/mundofesc/article/view/1198arquitectura de soluciónartefactos heterogéneosdeuda técnica en arquitecturamachine learning
spellingShingle Boris Rainiero Pérez-Gutiérrez
Darío Ernesto Correal
Fredy Humberto Vera-Rivera
Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura
Mundo Fesc
arquitectura de solución
artefactos heterogéneos
deuda técnica en arquitectura
machine learning
title Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura
title_full Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura
title_fullStr Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura
title_full_unstemmed Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura
title_short Un enfoque de machine learning para apoyar la identificación de la deuda técnica en arquitectura
title_sort un enfoque de machine learning para apoyar la identificacion de la deuda tecnica en arquitectura
topic arquitectura de solución
artefactos heterogéneos
deuda técnica en arquitectura
machine learning
url https://www.fesc.edu.co/Revistas/OJS/index.php/mundofesc/article/view/1198
work_keys_str_mv AT borisrainieroperezgutierrez unenfoquedemachinelearningparaapoyarlaidentificaciondeladeudatecnicaenarquitectura
AT darioernestocorreal unenfoquedemachinelearningparaapoyarlaidentificaciondeladeudatecnicaenarquitectura
AT fredyhumbertoverarivera unenfoquedemachinelearningparaapoyarlaidentificaciondeladeudatecnicaenarquitectura