Comparing Multivariate Time Series Analysis and Machine Learning Performance for Technical Debt Prediction: The SQALE Index Case
Robredo, Mikel; Saarimäki, Nyyti; Peñaloza, Rafael; Taibi, Davide; Lenarduzzi, Valentina (2024-06-07)
ACM
07.06.2024
Mikel Robredo, Nyyti Saarimaki, Rafael Penaloza, Davide Taibi, and Valentina Lenarduzzi. 2024. Comparing Multivariate Time Series Analysis and Machine Learning Performance for Technical Debt Prediction: The SQALE Index Case. In Proceedings of the 7th ACM/IEEE International Conference on Technical Debt (TechDebt '24). Association for Computing Machinery, New York, NY, USA, 45–46. https://doi.org/10.1145/3644384.3644472
https://creativecommons.org/licenses/by/4.0/
© 2024 Copyright held by the owner/author(s). This work licensed under Creative Commons Attribution International 4.0 License.
https://creativecommons.org/licenses/by/4.0/
© 2024 Copyright held by the owner/author(s). This work licensed under Creative Commons Attribution International 4.0 License.
https://creativecommons.org/licenses/by/4.0/
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:oulu-202408085275
https://urn.fi/URN:NBN:fi:oulu-202408085275
Tiivistelmä
Abstract
Predicting Technical Debt has become a popular research niche in recent software engineering literature. However, there is no consistent approach yet that succeeds in entirely capturing the nature of this type of data. We applied each technique on a dataset consisting of the commit data of a total of 28 Java projects. We predicted the future values of the SQALE index to evaluate their predictive performance. Using these techniques we confirmed the predictive power of each of them with the same commit data. We aim to investigate further the time-dependent nature of other types of commit data to validate the existing prediction techniques.
Predicting Technical Debt has become a popular research niche in recent software engineering literature. However, there is no consistent approach yet that succeeds in entirely capturing the nature of this type of data. We applied each technique on a dataset consisting of the commit data of a total of 28 Java projects. We predicted the future values of the SQALE index to evaluate their predictive performance. Using these techniques we confirmed the predictive power of each of them with the same commit data. We aim to investigate further the time-dependent nature of other types of commit data to validate the existing prediction techniques.
Kokoelmat
- Avoin saatavuus [34237]
Ellei muuten mainita, aineiston lisenssi on https://creativecommons.org/licenses/by/4.0/