lecture 22 23, intro decision making Flashcards
what uncertainties must we as engineers face
aleatory (natural randomness, weather etc)
epistemic (uncertainties in our data)
how do we deal with these uncertainties (3)
ignore,
us intuition and incorporate into models
use maths and CI intervals
type 1 and type 2 error
type 1, miscarrage of justice, (incorrectly reject null)
type 2 justice has not been served ( incorrectly accepted null)
what is the state of nature
it is the true value of the uncertain variable
for Bayes theorem , the decision in sensitive if
decision changes if tests done in different order
decision changes as tests are added
sensitivity will decrease as more tests are added
can choose not to do test if
it is thought that it is unlikely to change the dicsion
should only do a test if
the expected value gained is more than the cost
pre posterior analyisis involves
1 doing a test, chaning prob and deciding on action
2 finding prob given test already done
3 whether or not to do test anyway
WHAT IS ALLAIS paradox
people not making decisions on the basis of greatest utility
what is a risk neutral person
decision based on utility
what is a risk averse person
uncomfortable with small p*( dont want high chance of losing money)
what is a not risk averse person
will be comforatble with small p* ie comfortable with small prob ofr biug win.
who created the standard gamble
von neumann
how to do criteria for pure pessimism
identify smallest utlitity for each a, then choose largest of them.
how do, criteria of pure optimisim.
identify largest utility for each a, then choose largest of them.
