Please wait while the transcript is being prepared...
0:00
Welcome to this series of talks
entitled Expert Knowledge
Elicitation with SHELF.
This is the first
talk in the series,
it's entitled Uncertainty
and Probability, and my
name is Tony O'Hagan.
0:15
Here is the series
of talks as a whole.
As I said, there are four
talks in the series,
the first is this one;
Uncertainty and Probability,
where we'll talk
about elicitation,
which is a method of eliciting
probabilities essentially
to describe uncertainties.
The next two talks, the
second and the third,
are concerned with
real issues of
fundamentals of
the SHELF method,
which is the method of eliciting
expert knowledge that
we'll use in this course.
Finally, the fourth talk
is about techniques
and advanced things which
we'll come to when we get there.
0:48
The outline of this talk
is, there are two parts;
motivation and probability.
The first part talks about
uncertainty and decision-making,
and gives examples of where
expert knowledge
elicitation would
be important in decision-making.
Then we'll talk
about probabilities
and science, and
elicitation generally.
1:09
Motivation. Uncertainty
is everywhere.
1:12
It will always be with us,
it's always unwelcome;
people like certainty.
Admitting uncertainty is often
thought to be a
sign of weakness.
Politicians don't admit they're
uncertain about things.
People don't vote
for politicians that
say they're uncertain;
they want certainty.
Children are not
taught about it.
If you've got children,
you probably are not saying to
them when they ask you about
why something because there's
a certain probability
of something.
Children are not taught
about it at all, really.
They don't come to terms with
uncertainty except
when they're just
randomly faced with things
without ever quantifying
those uncertainties.
Uncertainty is the domain
for the statistician.
Statisticians work
with uncertainty.
We quantify it,
we model it, we analyse it,
and we do analysis of data
to reduce uncertainty.
Now for this talk, you
won't need to know
a lot about statistics because
elicitation is a
much simpler process
than all the
complicated statistics
that people are
taught at university.
