s*******r
发帖数: 769 | 1 就是把数据分成两部分，一部分60-70%，作为training data，剩下一部分作为
validaiton set。是叫 cross validation吗？ |
E**********e
发帖数: 1736 | 2 yes. there are also called in financial services as back testing or out of
time if your data are involved the time.
【在 s*******r 的大作中提到】
: 就是把数据分成两部分，一部分60-70%，作为training data，剩下一部分作为
: validaiton set。是叫 cross validation吗？
|
s*******r
发帖数: 769 | 3 谢谢
【在 E**********e 的大作中提到】
: yes. there are also called in financial services as back testing or out of
: time if your data are involved the time.
|
l******n
发帖数: 9344 | 4 Not out of time. It depends on whether training data has overlap with
validation data or not. It could be either in-time or our-of-time validation.
【在 E**********e 的大作中提到】
: yes. there are also called in financial services as back testing or out of
: time if your data are involved the time.
|
z**********e
发帖数: 91 | 5 不算吧，我理解是你要分成几组，然后每次拿其中的一组作为testing，剩下的作为
training,run多次，然后根据效果voting或者average。。。 |
