Analysis of Descriptors of Concept Drift and Their Impacts
Concept drift, a phenomenon that can lead to degradation of classifier performance over time, is commonly addressed in the literature through detection and reaction strategies. However, these strategies often rely on complete classifier retraining without considering the properties of the drift, whi...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Informatics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-9709/12/1/13 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850203735500783616 |
|---|---|
| author | Albert Costa Rafael Giusti Eulanda M. dos Santos |
| author_facet | Albert Costa Rafael Giusti Eulanda M. dos Santos |
| author_sort | Albert Costa |
| collection | DOAJ |
| description | Concept drift, a phenomenon that can lead to degradation of classifier performance over time, is commonly addressed in the literature through detection and reaction strategies. However, these strategies often rely on complete classifier retraining without considering the properties of the drift, which can prove inadequate in many scenarios. Limited attention has been given to understanding the nature of drift and its characterization, which are crucial for designing effective reaction strategies. Drift descriptors provide a means to explain how new concepts replace existing ones, offering valuable insights into the nature of drift. In this context, this work examines the impact of four descriptors—severity, recurrence, frequency, and speed—on concept drift through extensive theoretical and experimental analysis. Experiments were conducted on five datasets with 32 descriptor variations, eight drift detectors, and a non-detection context, resulting in 1440 combinations. The findings reveal three key conclusions: (i) reaction strategies must be tailored to different types of drift; (ii) severity, recurrence, and frequency descriptors have the highest impact, whereas speed has minimal influence; and (iii) there is a need to incorporate mechanisms for describing concept drift into the strategies designed to address it. |
| format | Article |
| id | doaj-art-683d4c6ebeeb4c2cb9457377cf667c5f |
| institution | OA Journals |
| issn | 2227-9709 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Informatics |
| spelling | doaj-art-683d4c6ebeeb4c2cb9457377cf667c5f2025-08-20T02:11:26ZengMDPI AGInformatics2227-97092025-01-011211310.3390/informatics12010013Analysis of Descriptors of Concept Drift and Their ImpactsAlbert Costa0Rafael Giusti1Eulanda M. dos Santos2Institute of Computing, Federal University of Amazonas, Av. Rodrigo Otávio, nº 6200, Coroado I, Campus Universitário Senador Arthur Virgílio Filho, Setor Norte, Manaus 69080-900, AM, BrazilInstitute of Computing, Federal University of Amazonas, Av. Rodrigo Otávio, nº 6200, Coroado I, Campus Universitário Senador Arthur Virgílio Filho, Setor Norte, Manaus 69080-900, AM, BrazilInstitute of Computing, Federal University of Amazonas, Av. Rodrigo Otávio, nº 6200, Coroado I, Campus Universitário Senador Arthur Virgílio Filho, Setor Norte, Manaus 69080-900, AM, BrazilConcept drift, a phenomenon that can lead to degradation of classifier performance over time, is commonly addressed in the literature through detection and reaction strategies. However, these strategies often rely on complete classifier retraining without considering the properties of the drift, which can prove inadequate in many scenarios. Limited attention has been given to understanding the nature of drift and its characterization, which are crucial for designing effective reaction strategies. Drift descriptors provide a means to explain how new concepts replace existing ones, offering valuable insights into the nature of drift. In this context, this work examines the impact of four descriptors—severity, recurrence, frequency, and speed—on concept drift through extensive theoretical and experimental analysis. Experiments were conducted on five datasets with 32 descriptor variations, eight drift detectors, and a non-detection context, resulting in 1440 combinations. The findings reveal three key conclusions: (i) reaction strategies must be tailored to different types of drift; (ii) severity, recurrence, and frequency descriptors have the highest impact, whereas speed has minimal influence; and (iii) there is a need to incorporate mechanisms for describing concept drift into the strategies designed to address it.https://www.mdpi.com/2227-9709/12/1/13concept driftdescriptorreaction strategyexplaining concept drift |
| spellingShingle | Albert Costa Rafael Giusti Eulanda M. dos Santos Analysis of Descriptors of Concept Drift and Their Impacts Informatics concept drift descriptor reaction strategy explaining concept drift |
| title | Analysis of Descriptors of Concept Drift and Their Impacts |
| title_full | Analysis of Descriptors of Concept Drift and Their Impacts |
| title_fullStr | Analysis of Descriptors of Concept Drift and Their Impacts |
| title_full_unstemmed | Analysis of Descriptors of Concept Drift and Their Impacts |
| title_short | Analysis of Descriptors of Concept Drift and Their Impacts |
| title_sort | analysis of descriptors of concept drift and their impacts |
| topic | concept drift descriptor reaction strategy explaining concept drift |
| url | https://www.mdpi.com/2227-9709/12/1/13 |
| work_keys_str_mv | AT albertcosta analysisofdescriptorsofconceptdriftandtheirimpacts AT rafaelgiusti analysisofdescriptorsofconceptdriftandtheirimpacts AT eulandamdossantos analysisofdescriptorsofconceptdriftandtheirimpacts |