|
- UID
- 223397
- 帖子
- 276
- 主题
- 21
- 注册时间
- 2011-7-11
- 最后登录
- 2014-8-7
|
3#
发表于 2011-7-13 13:34
| 只看该作者
Thanks. helps a lot.
what is multiple-R^2 ? Thanks.
MFIN--- Wrote:
-------------------------------------------------------
> R squared is the coefficient of determination that
> shows the total variation (not to be confused with
> variance) in the dependent variable explained by
> the independent variables. I.e., RSS divided by
> SST
>
> The problem with this measure is that it keeps
> increasing everytime you add another explanatory
> variable, even if it is not adding value to the
> model. To account for this issue, R squared
> adjusted is calculated which just to be noted can
> actually be negative in certain cases.
>
> R squared in general is very important in seeing
> whether a model is capturing the variation in what
> you're trying to explain! High R squared is
> typical for a good model (but it isn't always very
> high, depending on what type of series you're
> dealing with)
>
> Hope this helped |
|
