Learning from Imbalanced Data¶
Haibo He, & Garcia, E. A. (2009). Learning from Imbalanced Data. IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263-1284.
Cited by¶
2 citations across 2 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- False Positive Paradox
- In machine learning at scale, class imbalance is the paradox in algorithmic dress, and practitioners reach for precision, recall, and PR-AUC precisely to surface it.
This sourceSurveys class-imbalance methods — precision/recall, PR curves, cost-sensitive learning, threshold tuning — that surface and address the rare-positive regime, supporting the machine-learning use and the ML-to-ecology toolbox transfer.
Supported in partVerified against the publisher's abstract
The abstract confirms this review treats class imbalance as a metrics problem, reviewing assessment metrics for skewed data, but it names no specific metric and says nothing about forensic statistics.
“The imbalanced learning problem is concerned with the performance of learning algorithms in the presence of underrepresented data and severe class distribution skews. ... Our focus is to provide a critical review of the nature of the problem, the state-of-the-art technologies, and the current assessment metrics used to evaluate learning performance under the imbalanced learning scenario.”
- In machine learning at scale, class imbalance is the paradox in algorithmic dress, and practitioners reach for precision, recall, and PR-AUC precisely to surface it.
- Majority-Dominated Aggregate Objective
- Machine learning — aggregate cross-entropy on an imbalanced dataset is dominated by majority-class terms; the classifier learns "predict majority" as a low-loss strategy and fails on the minority the application cares about (diagnosis, fraud, rare-defect detection).
This sourceSurveys class imbalance: an aggregate loss dominated by the majority class makes 'predict majority' a low-loss strategy with near-zero minority recall, with reweighting, resampling, and threshold-tuning as remedies.
Supported in partVerified against the publisher's abstract
The survey's abstract establishes the imbalanced-learning problem (under-represented data, skewed class distributions, degraded learning performance) but not the claim's loss-domination mechanism or its application examples.
“The imbalanced learning problem is concerned with the performance of learning algorithms in the presence of underrepresented data and severe class distribution skews.”
- Machine learning — aggregate cross-entropy on an imbalanced dataset is dominated by majority-class terms; the classifier learns "predict majority" as a low-loss strategy and fails on the minority the application cares about (diagnosis, fraud, rare-defect detection).
Verification¶
Does it exist? Confirmed. This work's DOI resolves to a registered record, which fixes its identity. That is all it fixes.
Does it back the claim? Read against the text for 2 of 2 citations: 2 supported in part. Each verdict is shown under its citation below, with what in the work backs the sentence.
Was it audited? Yes. A second, independent pass read the citation against the article text and recorded a verdict.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Registry ID ref:4fd0ed59350a · see in the full table