Area under the Precision-Recall Curve: Point Estimates and Confidence Intervals
By Kendrick Boyd, Kevin H. Eng, David Page. ECML/PKDD, 2013
525 citations recorded by the source corpus.
Subfields: Artificial Intelligence
View Paper: https://www.semanticscholar.org/paper/7baa4f71cff73f744ad745735bc3af1c20ed4ced
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Aaron C. Courville
- Faruk Ahmed
- Yoshua Bengio
- Woong Bae
- Seyed Sadegh Mohseni Salehi
- Jayashree Kalpathy-Cramer
- A. Gholipour
- Julius Adebayo
Citations in the corpus
- Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMs (2023, methodology): “, 2022), we use confidence calibration and failure prediction metrics to measure estimated confidence: Expected Calibration Error (ECE) that measures the discrepancy between predicted probabilities and observed accuracy within each confidence level; Area Under the Receiver Operating Characteristic c”
- Assessing the (Un)Trustworthiness of Saliency Maps for Localizing Abnormalities in Medical Imaging (2020, methodology): “…boxes in the training and validation datasets, and a high baseline determined by the Area Under the Precision Recall Curve (AUPRC [20] 1 )of segmentation (U-Net) and detection networks (RetinaNet); b) in model weight randomization, the average Structural SIMilarity (SSIM) 2…”
- Detecting semantic anomalies (2019, methodology): “Interpolation in a precision-recall curve can sometimes be misleading, as studied in Boyd, Eng, and Page (2013), who recommend a number of more robust estimators.”
- Tversky Loss Function for Image Segmentation Using 3D Fully Convolutional Deep Networks (2017, methodology): “Compared to DSC which weighs recall and precision equally, and the ROC analysis, we consider the area under the PR curves (APR, shown in Figure 2) the most reliable performance metric for such highly skewed data [8,1].”
- IDRiD: Diabetic Retinopathy - Segmentation and Grading Challenge. (2020, methodology): “The AUPR measure is more realistic ( Boyd et al., 2013; Saito and Rehmsmeier, 2015 ) for the lesion segmentation performance over the area under Receiver Operating Characteristic (ROC) curve.”