### Abstract

Maximum likelihood estimation of receiver operating characteristic (ROC) curves using the "proper" binormal model can be interpreted in terms of Bayesian estimation as assuming a flat joint prior distribution on the c and d _{a} parameters However, this is equivalent to assuming a non-flat prior distribution for the area under the curve (AUC) that peaks at AUC - 1.0. We hypothesize that this implicit prior on AUC biases the maximum likelihood estimate (MLE) of AUC We propose a Bayesian implementation of the "proper" binomial ROC curve-fitting model with a prior distribution that is marginally flat on AUC and conditionally flat over c. This specifies a non-flat joint prior for c and d _{a} We developed a Monte Carlo Markov chain (MCMC) algorithm to estimate the posterior distribution and the maximum a posteriori (MAP) estimate of AUC. We performed a simulation study using 500 draws of a small dataset (25 normal and 25 abnormal cases) with an underlying AUC value of 0.85. When the prior distribution was a flat joint prior on c and d _{a} the MLE and MAP estimates agreed, suggesting that the MCMC algorithm worked correctly. When the prior distribution was marginally flat on AUC, the MAP estimate of AUC appeared to be biased low. However the MAP estimate of AUC for perfectly separable degenerate datasets did not appear to be biased. Further work is needed to validate the algorithm and refine the prior assumptions.

Original language | English (US) |
---|---|

Title of host publication | Medical Imaging 2007 |

Subtitle of host publication | Image Perception, Observer Performance, and Technology Assessment |

DOIs | |

State | Published - Oct 15 2007 |

Event | Medical Imaging 2007: Image Perception, Observer Performance, and Technology Assessment - San Diego, CA, United States Duration: Feb 21 2007 → Feb 22 2007 |

### Publication series

Name | Progress in Biomedical Optics and Imaging - Proceedings of SPIE |
---|---|

Volume | 6515 |

ISSN (Print) | 1605-7422 |

### Conference

Conference | Medical Imaging 2007: Image Perception, Observer Performance, and Technology Assessment |
---|---|

Country | United States |

City | San Diego, CA |

Period | 2/21/07 → 2/22/07 |

### Fingerprint

### Keywords

- "Proper" receiver operating characteristic curve analysis
- Bayesian estimation
- Monte Carlo Markov chains

### ASJC Scopus subject areas

- Electronic, Optical and Magnetic Materials
- Atomic and Molecular Physics, and Optics
- Biomaterials
- Radiology Nuclear Medicine and imaging

### Cite this

*Medical Imaging 2007: Image Perception, Observer Performance, and Technology Assessment*[651511] (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 6515). https://doi.org/10.1117/12.711689

}

*Medical Imaging 2007: Image Perception, Observer Performance, and Technology Assessment.*, 651511, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 6515, Medical Imaging 2007: Image Perception, Observer Performance, and Technology Assessment, San Diego, CA, United States, 2/21/07. https://doi.org/10.1117/12.711689

**A Bayesian interpretation of the "proper" binormal ROC model using a uniform prior distribution for the area under the curve.** / Zur, Richard M.; Pesce, Lorenzo Luigi; Jiang, Yulei; Metz, Charles E.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - A Bayesian interpretation of the "proper" binormal ROC model using a uniform prior distribution for the area under the curve

AU - Zur, Richard M.

AU - Pesce, Lorenzo Luigi

AU - Jiang, Yulei

AU - Metz, Charles E.

PY - 2007/10/15

Y1 - 2007/10/15

N2 - Maximum likelihood estimation of receiver operating characteristic (ROC) curves using the "proper" binormal model can be interpreted in terms of Bayesian estimation as assuming a flat joint prior distribution on the c and d a parameters However, this is equivalent to assuming a non-flat prior distribution for the area under the curve (AUC) that peaks at AUC - 1.0. We hypothesize that this implicit prior on AUC biases the maximum likelihood estimate (MLE) of AUC We propose a Bayesian implementation of the "proper" binomial ROC curve-fitting model with a prior distribution that is marginally flat on AUC and conditionally flat over c. This specifies a non-flat joint prior for c and d a We developed a Monte Carlo Markov chain (MCMC) algorithm to estimate the posterior distribution and the maximum a posteriori (MAP) estimate of AUC. We performed a simulation study using 500 draws of a small dataset (25 normal and 25 abnormal cases) with an underlying AUC value of 0.85. When the prior distribution was a flat joint prior on c and d a the MLE and MAP estimates agreed, suggesting that the MCMC algorithm worked correctly. When the prior distribution was marginally flat on AUC, the MAP estimate of AUC appeared to be biased low. However the MAP estimate of AUC for perfectly separable degenerate datasets did not appear to be biased. Further work is needed to validate the algorithm and refine the prior assumptions.

AB - Maximum likelihood estimation of receiver operating characteristic (ROC) curves using the "proper" binormal model can be interpreted in terms of Bayesian estimation as assuming a flat joint prior distribution on the c and d a parameters However, this is equivalent to assuming a non-flat prior distribution for the area under the curve (AUC) that peaks at AUC - 1.0. We hypothesize that this implicit prior on AUC biases the maximum likelihood estimate (MLE) of AUC We propose a Bayesian implementation of the "proper" binomial ROC curve-fitting model with a prior distribution that is marginally flat on AUC and conditionally flat over c. This specifies a non-flat joint prior for c and d a We developed a Monte Carlo Markov chain (MCMC) algorithm to estimate the posterior distribution and the maximum a posteriori (MAP) estimate of AUC. We performed a simulation study using 500 draws of a small dataset (25 normal and 25 abnormal cases) with an underlying AUC value of 0.85. When the prior distribution was a flat joint prior on c and d a the MLE and MAP estimates agreed, suggesting that the MCMC algorithm worked correctly. When the prior distribution was marginally flat on AUC, the MAP estimate of AUC appeared to be biased low. However the MAP estimate of AUC for perfectly separable degenerate datasets did not appear to be biased. Further work is needed to validate the algorithm and refine the prior assumptions.

KW - "Proper" receiver operating characteristic curve analysis

KW - Bayesian estimation

KW - Monte Carlo Markov chains

UR - http://www.scopus.com/inward/record.url?scp=35148845650&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35148845650&partnerID=8YFLogxK

U2 - 10.1117/12.711689

DO - 10.1117/12.711689

M3 - Conference contribution

SN - 0819466336

SN - 9780819466334

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2007

ER -