Hi SAS community members,
I need to find a cut-off value for Estimated Probability that make EC has the minimum value
EC= q1(x1)c1+q2(x2)c2
where q1=0.01, c1=1 , q2=0.99, c2=20
x1 is the probability of observed Type I error , x2 is the probability of observed the Type II error
q1 is the assuemd probabilty for manipulator = 0.01 (1%), and the assumend probabilty for non-manipulator is q2=0.99 (99%).
The following is the sample data
Global Company | em | Estimated Probability |
Key |
|
|
1173 | 1 | 0.01095 |
1173 | 1 | 0.00952 |
1286 | 1 | 0.00804 |
3362 | 1 | 0.0161 |
3734 | 0 | 0.00812 |
3839 | 1 | 0.0123 |
3946 | 1 | 0.01109 |
4060 | 0 | 0.00972 |
4060 | 1 | 0.01077 |
4372 | 1 | 0.00967 |
4601 | 1 | 0.00978 |
4601 | 0 | 0.00981 |
4622 | 1 | 0.00922 |
4769 | 1 | 0.01033 |
5047 | 1 | 0.01023 |
5399 | 1 | 0.00945 |
6136 | 1 | 0.01154 |
6136 | 1 | 0.00984 |
6882 | 1 | 0.00917 |
7652 | 1 | 0.00949 |
7652 | 0 | 0.00958 |
7991 | 1 | 0.01051 |
8092 | 1 | 0.01075 |
8972 | 0 | 0.01 |
9538 | 1 | 0.01213 |
10026 | 1 | 0.00946 |
em=1 means the observations is an earnings manipulator, 0 otherwise,
Estimated Probability is the estmated probabilty calculated by a Logit model.
In my sample, there are 100 manipulators and 9,900 non-manipulators.
Could anyone please help me for this .
Many thanks for your help.