A*******s
发帖数: 3942 | 1 dont think so... even though SAS/IML and SAS/OR have quadratic programming
functions/procedures, implementing SVM is not trivial anyway in SAS. |
|
s**f
发帖数: 365 | 2 对于continuous的variable，都加上至少quadratic和cubic的term，然后test是不是
significant |
|
F****n
发帖数: 3271 | 3 Discretize and / or test quadratic term. It's technically simple but kind of testing experience. The non-linearity of age should be always be addressed, especially this is about smoking.
For example, think about children / adult. Even the data are all adults, non-linear effects can also occur. |
|
z**********i
发帖数: 12276 | 4 有个SURVEY的DATA, 10-12岁,10个国家.
LEVEL 1 VARIABLE(STUDENT):
BMI(BODY MASS INDEX, OUTCOME VARIABLE), AGE, GENDER, PHYSICAL ACTIVITY,
DIETARY
LEVEL 2 VARIABLE(COUNTRY):
INCOME, COUNTRY NAME
问题:
想体现多个(10)国家的MULTI-CULTURE,有什么常用的VARIABLE?如果用COUNTRY NAME,是
不是LEVEL 太多了? 如果用LANGUAGE,基本上每过国家都有自己的LANGUAGE. 用洲来分
呢?亚洲,美洲,非洲?
第一次用这个REGRESSION,能给些建议吗?如何考虑加入QUADRATIC TERM 或INTERACTION
TERM?或者应该读些什么文章?
多谢!! |
|
z**********i
发帖数: 12276 | 5 有个SURVEY的DATA, 10-12岁,10个国家.
LEVEL 1 VARIABLE(STUDENT):
BMI(BODY MASS INDEX, OUTCOME VARIABLE), AGE, GENDER, PHYSICAL ACTIVITY,
DIETARY
LEVEL 2 VARIABLE(COUNTRY):
INCOME, COUNTRY NAME
问题:
想体现多个(10)国家的MULTI-CULTURE,有什么常用的VARIABLE?如果用COUNTRY NAME,是
不是LEVEL 太多了? 如果用LANGUAGE,基本上每过国家都有自己的LANGUAGE. 用洲来分
呢?亚洲,美洲,非洲?
第一次用这个REGRESSION,能给些建议吗?如何考虑加入QUADRATIC TERM 或INTERACTION
TERM?或者应该读些什么文章?
多谢!! |
|
l*********s
发帖数: 5409 | 6 Say, a regressor X enters into the regression model in the quadratic form
a*(X-b)^2 ,
is there some measure similar to R square in linear models that reflect the contribution of X to the explained variance of dependent variable? Thanks a bunch. |
|
z*******n
发帖数: 15481 | 7 这是quadratic form
发我十个包子 完整告诉你下面该做啥 |
|
A*******s
发帖数: 3942 | 8 google "bounded outcome regression". As far as i know, latent variable model
or beta regression can be applied.
i dont think it is about lack of curvature. I bet LZ would see the same
pattern even after he add quadratic terms.
it is ok to use OLS in the case if only the conditional mean of Y is of
interest and variance is nuisance. Whenever u would like to draw inference
from the distribution of Y (significance test, confidence interval), OLS
would fail since it gives wrong estimate of VAR[Y|X] ... 阅读全帖 |
|
z*******n
发帖数: 15481 | 9 晕菜 那几乎所有问题都是bounded的了 你凭什么假设最小的p是0.01.。。 如果是0.
001 那下线会降低多少
有意思么。。。
还非得当bounded来做。。。。。你统计白学了啊
你那个是0 1 bounded当然不一样
这个就是普通的问题 加入quadratic term就可以很有效了 根本都不需要
transformation |
|
A*******s
发帖数: 3942 | 10 BTW u need to read LZ's post again. The residual plot is "response vs
residual" rather than "residual vs response". Please turn ur head 90 degree
before u suggest adding quadratic term. |
|
A*******s
发帖数: 3942 | 11 sas的优势还是大而全，
举个例子，找一个能覆盖glimmix全部功能的软件包不容易。
我没在it公司工作过，不好说你们用啥模型。
不过我不明白，在业界搞data mining应用的(非研究)，
有多少需要hard core programming skills呢？
难道说搞一个regression都要自己从头写SVD？
搞一个SVM都得自己从头搞quadratic programming？
搞一个RF都得自己从头搞recursive partition+parallel？
如果只是调用别人写好的包的话
从难度上来说，和用SAS Procedure有啥区别？ |
|
|
s*r
发帖数: 2757 | 13 would a quadratic term be helpful |
|
n**m
发帖数: 156 | 14 After I ran the model, it reported:
WARNING: Obtaining minimum variance quadratic unbiased estimates as starting
values for the
covariance parameters failed.
不知道有没有什么办法解决这个问题？ |
|
q**j
发帖数: 10612 | 15 就是能够解决quadratic programming的。有人从python里面叫R里面的R metrics么？
这个应该可以，但是有点担心是否能够handle大数据和多变量（一次可以解1万个变量
的那种。）多谢。 |
|
q**j
发帖数: 10612 | 16 就是能够解决quadratic programming的。有人从python里面叫R里面的R metrics么？
这个应该可以，但是有点担心是否能够handle大数据和多变量（一次可以解1万个变量
的那种。）多谢。 |
|
o****o
发帖数: 8077 | 17 build a quadratic random growth curve, subject=Sub |
|
y****2
发帖数: 46 | 18 I have a dataset which fits the non-linear regression. But which parameter
should I use to judge the goodness of fit? something like R square.
The non-linear equation is second order polynomial(quadratic) function.
Thank you everyone. |
|
l**********8
发帖数: 305 | 19 for ordinal or continuous variables, may also determine if the
ln(odds) seems to be linearly related or not. Plot the logit of the group
means
vs the midpoint of the group. Will give an idea of whether to model as
linear, quadratic, or possibly just categorical |
|
l******t
发帖数: 96 | 20 你举来举去除了一个t test还有啥？
任何学科开端从不会一上来就做最高深的东西。那我还可以说数学一开始就是算术，有
啥难的？
Neyman-Pearson，UMP， quadratic mean differentiability一样用了很多分析的知识 |
|
c***z
发帖数: 6348 | 21 Andrew Ng's machine learning course on coursera can be a good start
at the end of day, it is all about minimizing a cost function, usually a
quadratic one... |
|
d******e
发帖数: 7844 | 22 Newton和QN的收敛都和Hessian的condition number有关。
QN是superlinear，虽然不如Newton的quadratic，但是每个iteration的计算复杂度通
常至少节省一个problem dimension的factor。所以最后通常比Newton快。
比如L-BFGS，很多时候在大规模数据上的表现远远好过SGD和Newton。 |
|
d******e
发帖数: 7844 | 23 Newton和QN的收敛都和Hessian的condition number有关。
QN是superlinear，虽然不如Newton的quadratic，但是每个iteration的计算复杂度通
常至少节省一个problem dimension的factor。所以最后通常比Newton快。
比如L-BFGS，很多时候在大规模数据上的表现远远好过SGD和Newton。 |
|
i**y
发帖数: 1170 | 24 天朝子民只要美国货就fan的不行，没必要往亏里搞，很多东西原价货他们都觉得便宜
大公司诸如amazon，macys啥的就没戏了，小公司例如DSCM什么的，还是很有搞头的，
这次俺就猜到它revenue一定会beat to da moon （野鸡订单号回归发现quadratic
term超显著，thanksgiving连出几个outlier，大概算了一下花街的109M sales明显不
靠谱），于是赌涨，蒙对了，从2.7一路涨到3.7.
但其实谁知道呢，没准花街认为revenue beat也没啥，毕竟它还是亏损的。 |
|
