s*********e
发帖数: 1051 | 1 if you are interested in consumer scorecard, you might want to take a look.
http://statcompute.wordpress.com/2013/04/27/a-sas-macro-calcula |
j**********3
发帖数: 305 | 2 thx wensui very neat code
【在 s*********e 的大作中提到】
: if you are interested in consumer scorecard, you might want to take a look.
: http://statcompute.wordpress.com/2013/04/27/a-sas-macro-calcula
|
s******0
发帖数: 1269 | |
s*********e
发帖数: 1051 | 4 coding is nothing. it is the idea and practice.
【在 j**********3 的大作中提到】
: thx wensui very neat code
|
j**********3
发帖数: 305 | 5 It was a very useful tool for score alignment of ongoing scorecard
monitoring and support. ensuring the same score represents the same actual
odds across all models in a multi-model environment is critical !
Thanks man ! |
t**********r
发帖数: 98 | 6 thanks for sharing, by the way, what is the trigger to decide the model is
not accuracy/or model's PDO is not acceptble if my development PDO=20? |
A*******s
发帖数: 3942 | 7 大侠这个macro我用过，非常不错。不过有个小feedback，就是里面的对score和weight
的roundoff会带来一些误差，我觉得也没有必要。 |
s*********e
发帖数: 1051 | 8 valid point
weight
【在 A*******s 的大作中提到】
: 大侠这个macro我用过，非常不错。不过有个小feedback，就是里面的对score和weight
: 的roundoff会带来一些误差，我觉得也没有必要。
|
