Title: Presentations
1Presentations
- For presentations Bring 3 copies of slides, two
to turn in (4 slides per page) and one copy in
case of computer problems (1 slide per page) - Check website for general guidelines on
presentations and papers - Check website for due dates on papers.
- Bring to class for presentationsCopies of peer
evaluation form, 6 forms, one for each
presentation other than your own. These will be
due each day following the presentations.
2Scenario
- You work at a bank and need to decide whether to
foreclose on a loan for a borrower who is
delinquent - Arcadia Bank has a set of records from 3
banks1. BR bank has records of of years
experience2. Cajun Bank has records of
educational level3. DuPont bank has records on
state of the economy - NOTE Each bank has records only regarding ONE
criteria and whether the loan was paid off or not
3Data Notation
- The data set is a collection of successes and
failures for loan workouts, divided into three
categories, Y, T, C - Notation and variablesSthe set of all
successful loan workoutsFthe set of all failed
workoutsYthe event that a borrower has y years
experienceTthe event that a borrower has
educational level tCthe event that a borrower
has economic condition c - NOTE The information given on each of the
borrowers satisfies one and only one of the
events Y, T, or C.
4Goal
- The goal is to use all available information to
compute the expected return on your borrowers
loan workout - Let Zthe random variable giving the amount of
money that Arcadia Bank will receive from a loan
workout - Z can take on two values1. full value if loan
is successful2. default value if the loan is a
failure
5Expected Return
- For your preliminary reports you
calculatedE(Z)full valueP(S)default
valueP(F) - This was just a simple calculation and did not
take into account the specific data on your
borrower - Now we can use all the tools we have learned in
order to make a quantitative decision on whether
to do a workout or foreclosure on your borrower - Why do we need to make a quantitative decision?
Address this in your project.
6Expected Value Using Y,T,C
- E(Z)full P(SYnTnC) default P(FYnTnC)
Lets look at P(SYnTnC)
(By definition of conditional probability)
(By definition of Bayes Theorem)
7Expected Value Using Y,T,C, cont
This is true because Y, T and C are independent
events, which implies that
are independent events
Since they are independent, we can multiply their
probabilities You need to explain why they are
independent.
8Expected Value Using Y,T,C, cont
- We can closely estimate the value of
by using our database
9E(ZYTC)
- E(ZYTC)full P(SYnTnC) default P(FYnTnC)
10E(ZYTC), cont
So you must calculate this as part of your
decision making process. When you talk about
this calculation, do not beat the point to
death!! In other words, do not tell me step by
step what you did (i.e, multiply by this number
and add it to this number, etc.)
11E(ZYTC), cont
- We just calculated part of our E(zYTC) value
- E(Z)full P(SYnTnC) default P(FYnTnC)
- NOTE P(SYnTnC) P(FYnTnC)1
- Why?
12E(ZYTC)
- Just when you thought it was all over
- Now you go through the calculation again except
this time, using Y and Y - EX if Y4 yearsY3-5 years ( a range of
years)Y2-6 years
13E(ZYTC)
- Use the values of E(ZYTC), E(ZYTC), E(ZYTC) to
make your decision - NOTE that using Y or Y does NOT require a
totally new calculation. Only the probabilities
dealing with the Y change, everything else stays
the same
14Things to consider
- Why do we use Expected Value to make our
decision? - Why are Y, T and C independent?
- Where is independence used?
- Why do you use Bayes Theorem?
- Why can you use Bayes Theorem?
15Further Analysis
- The entire project has been laid out for you
you should now know exactly what to do so you can
make a decision - I want you to do some further analysis on this
project - Things to consider What would happen if the
economy changed? Education level? - Make your project be unique and stand out!