Abstract
There are several important advantages in applying conformal predictors (CP) in medical diagnostics. First of all, CPs provide valid measures of confidence in the diagnosis; this is often a crucial advantage for medical decision-making since it allows the estimation of risk of an erroneous clinical decision for an individual patient. Moreover, the risk of clinical errors may be controlled by an acceptable level of confidence for a given clinical decision and therefore the risk of misdiagnosis is known. Another feature that makes CPs an attractive method in medical applications is that they are region predictors. This means that if we do not have enough information to make a definitive diagnosis, the method would allow us to make a number of possible (multiple) diagnoses and a patient may require further tests to narrow down the available options.
Original language | English |
---|---|
Title of host publication | Conformal Prediction for Reliable Machine Learning |
Subtitle of host publication | Theory, Adaptations and Applications |
Publisher | Elsevier Inc. |
Pages | 217-230 |
Number of pages | 14 |
ISBN (Print) | 9780123985378 |
DOIs | |
Publication status | Published - Apr 2014 |
Externally published | Yes |
Keywords
- Biomedical Applications
- Diagnostics
- Microarrays
- Nonconformity Measures
- Prognostics
- Proteomics
ASJC Scopus subject areas
- General Computer Science