Statistical methods for analyzing binary replicates, i.e., noisy binary measurements of latent binary states. This package implements the methods described in:
Royer-Carenzi, M., Lorenzo, H., & Pudlo, P. (in press). Reconciling Binary Replicates: Beyond the Average. Statistics in Medicine.
Overview
The package provides scoring functions to estimate the probability that an individual is in the positive state, given noisy replicated measurements:
| Method | Function | Requirements |
|---|---|---|
| Average-based | average_scoring() |
None |
| Median-based | median_scoring() |
None |
| MAP (EM algorithm) | MAP_scoring() |
Fitted EM model |
| Likelihood-based | likelihood_scoring() |
Known parameters |
| Bayesian | bayesian_scoring() |
Fitted Bayesian model |
Additional features:
- Classification with inconclusive decisions (
classify_with_scores()) - Prevalence estimation from scores or Bayesian posterior
- Credible intervals for model parameters
- Cross-validation for model assessment (
cvEM())
Statistical Model
For each individual