expert: Modeling Without Data Using Expert Opinion
Expert opinion is a technique to do statistical modeling when data is
scarse (e.g. accidents in nuclear plants) or even absent, at least for
the analyst (e.g. confidential settlements in liability insurance).
Opinions on the distribution of the decision random variable is sought
from experts in the field. The experts give their opinion in the form
of a few quantiles for the decision variable and for a set of "seed
variables" for which the analyst knows the true quantiles. Results for
seed variables are compared to the true values and used to determine
the influence of each expert on the aggregated distribution. The
package supports three different ways to aggregate the information
provided by the experts in one final distribution:
1. the classical model of Cooke (1991);
2. the Bayesian model of Mendel and Sheridan (1989);
3. an ad-hoc procedure where the weight of each expert is pre-
determined by the analyst.
The main function of the package is expert(), a unified interface to
all three methods above. The package also provides a few utility
functions to display, plot or compute probabilities and quantiles from
the aggregated distribution returned by expert().
Best regards,
Mathieu Pigeon
Institut de Statistique
Universite Catholique de Louvain
Voie du Roman Pays, 20
1348 Louvain-la-Neuve
BELGIUM
E-mail address : mathieu.pigeon at uclouvain.be
_______________________________________________
R-packages mailing list
R-packages at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-packages
