Entropy-based dynamic ensemble classication algorithm for imbalanced data stream with concept drift

JiaMing Gong, MingGang Dong, Suja A Alex, Hugh Cowley, Unil Yun, Unil Yun, Unil Yun · 2024

Read Entropy-based dynamic ensemble classication algorithm for imbalanced data stream with concept drift by JiaMing Gong, MingGang Dong, Suja A Alex, Hugh Cowley, Unil Yun, Unil Yun, Unil Yun (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11643253.txt.

Description

The authors present a learning algorithm in a 2024 paper for data that arrives as an endless stream. Two headaches come with streams: the classes can be badly lopsided, with rare cases drowned by common ones, and the pattern itself can drift over time.

Their algorithm builds a team of classifiers and reweights them as fresh chunks of data land. To fight lopsidedness, it splits each chunk into balanced pairs guided by an entropy measure, and a density check picks the best rare-case examples for training. A feedback loop sets each team member's vote weight from its recent form. In tests against five rival methods, the new algorithm came out ahead on standard stream benchmarks.

Catalog record

CREATORJiaMing Gong, MingGang Dong, Suja A Alex, Hugh Cowley, Unil Yun, Unil Yun, Unil Yun
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE96 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC11643253

Browse the full library · About the Tower of Babel