p****x 发帖数: 1346 | 1 If I have all the consumer loans history data ready, how to create
Probability Default rate & Loss Given Default rate?
Thanks! | s**********l 发帖数: 395 | 2 #1. To create PD:
After processing and cleaning the data, choose model(usually logistic
regression)and input variables, then you can build your model.After that,
you can get PD.
#2. To create LGD: similar to PD, except using the general linear regression
model. | p****x 发帖数: 1346 | 3 还是不太明白, 能否给我举个例子? 我现在在准备timing curve, 他们之间是什么关
系呢?
regression
【在 s**********l 的大作中提到】 : #1. To create PD: : After processing and cleaning the data, choose model(usually logistic : regression)and input variables, then you can build your model.After that, : you can get PD. : #2. To create LGD: similar to PD, except using the general linear regression : model.
| s*********e 发帖数: 1051 | 4 misleading in 2
regression
【在 s**********l 的大作中提到】 : #1. To create PD: : After processing and cleaning the data, choose model(usually logistic : regression)and input variables, then you can build your model.After that, : you can get PD. : #2. To create LGD: similar to PD, except using the general linear regression : model.
| s**********l 发帖数: 395 | 5 Hi, Statcompute
Could you tell me why #2 is misleading? Then how to do LGD? Thanks.
【在 s*********e 的大作中提到】 : misleading in 2 : : regression
| s*********e 发帖数: 1051 | 6 since lgd = 1-recovery, it should be bounded between zero and one.
【在 s**********l 的大作中提到】 : Hi, Statcompute : Could you tell me why #2 is misleading? Then how to do LGD? Thanks.
|
|