Please wait while the transcript is being prepared...
0:00
Welcome to this third
talk in the series,
'Expert Knowledge
Elicitation with SHELF.'
This one is entitled
Multiple Experts.
0:11
There again is the
rundown of the course
and we are in the third talk
here, multiple experts.
In the last talk, number two,
we did one distribution
with one expert,
and now we're going to move
onto multiple experts,
which is the major area
where we actually use SHELF.
The final talk will be
some details of skills and
resources that we might do.
0:34
But before we do so,
let's look back at
part 2. In part 2 -
One Distribution,
One Expert - we found out about
challenges for subjective
probability judgements,
primarily in the form of
psychological heuristics that
can lead to biased judgements.
We saw how SHELF protocol
addresses those challenges.
We have a carefully
constructed sequence of
questions designed to minimise
the potential for bias.
We did that for one expert,
and now we're going to move
on to multiple experts.
1:05
So in part 3, we'll consider
the general problem of
eliciting a single distribution
from multiple experts.
We have to consider then how
to aggregate the experts
together using
either mathematical or
behavioural aggregation.
We'll see how SHELF uses
behavioural aggregation
to do this with
some specific new concepts
like the rational
impartial observer,
we'll deal with those.
Then we'll compare
the various SHELF
methods for eliciting
judgements from individual
experts and from the group.
1:36
This part will have
three parts to it,
the aggregation part,
multiple experts and
the SHELF methods.
