.png){kind=icon}
Return
Assessing machine learning model reliability using data-driven rules
.png){kind=icon}
OA
AI
DOI：10.1016/j.ress.2026.113144.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A data-driven rule framework is proposed to assess Machine Learning (ML) prediction reliability. • Pointwise reliability scores are derived from rule consistency and data density. • The method is model-agnostic and applicable to black-box predictors. • Clinical case studies show improved decision confidence in low-risk regions.
Keywords:
Machine learning models
GRACE risk score
Reliability metrics
Boolean logic
Rule-based approach
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
R
IF:
11
Papers:
1.0K
Citations:
0
Organization
Cited Papers
No Longer Confidential: Estimating the Confidence of Individual Regression Predictions.png)
PLoS ONEOAAI
PLoS ONEIF0
OAAIShareSaveEstimation of individual prediction reliability using the local sensitivity analysisAPPLIED INTELLIGENCE
OAAI
APPLIED INTELLIGENCEIF3.5
OAAIShareSaveShareSaveShareSaveThe role of AI in detecting and mitigating human errors in safety-critical industries: A review
PREAI
ShareSave
View more