Building models with SAS Enterprise Miner, SAS Factory Miner, SAS Visual Data Mining and Machine Learning or just with programming

Significance IV with a log transformed Dependent Variable

Accepted Solution Solved
Reply
Frequent Contributor
Posts: 92
Accepted Solution

Significance IV with a log transformed Dependent Variable

Hi All, Good Morning.

 

In a linear regression model that i am working on i have log transformed dependent variable due to data being skewed so below is the output from the model (Some of the values are mocked up with artificial var names):

 

Model

                       Parameter Estimate   Standard Error    Significance Value (P)

Intercept         7.346                                0.12                    .000

Carat              1.392                                .009                    .000

V GOOD        -.211                                 .016                    .000

GOOD           -.134                                 .013                     .000

Carat_I           -0.44                                .009                     .000

 

Here as the Dep var is log transformed we need to measure and interpret the coefficient in terms of percent change. So my question is while keeping var Carat as controlled which is most significant variable.

As per the estimates value it should be Good as it has a better slope value, however 

 

 the regression equation = 7.346-0.211 = 7.135 (with  V Good var)

 the regression equation = 7.346-0.134 = 7.212 (with Good var)

This has lead to some confusion. request you to please let me know if i am ok interpreting the results.

 

Regards, Shivi

 

 

 

 


Accepted Solutions
Solution
‎11-29-2015 07:49 AM
Frequent Contributor
Posts: 108

Re: Significance IV with a log transformed Dependent Variable

I am trying to answer without see all the parameters .You confusion is why V Good var is less then Good var . I belive it is because of the Std Error.You have classified VGV and GV due to some assumption or result , if those assumption or test are correct you are Good.

View solution in original post


All Replies
Solution
‎11-29-2015 07:49 AM
Frequent Contributor
Posts: 108

Re: Significance IV with a log transformed Dependent Variable

I am trying to answer without see all the parameters .You confusion is why V Good var is less then Good var . I belive it is because of the Std Error.You have classified VGV and GV due to some assumption or result , if those assumption or test are correct you are Good.
Frequent Contributor
Posts: 92

Re: Significance IV with a log transformed Dependent Variable

Thank you, yes i have checked all the assumptions and they are true. 

However if you check the regression equation the variable Good has better value then V Good and that is what is confusing here. 

Frequent Contributor
Posts: 92

Re: Significance IV with a log transformed Dependent Variable

Please ignore the below question of "However if you check the regression equation the variable Good has better value then V Good and that is what is confusing here"

 

After looking carefully i can relate why good will be than Vgood because these var are impacting the outcome negatively and good is impacting more than v good.

 

Sometimes you tend to get confused when you work closely with the people who are either new or are only masters on books.

 

Thanks for the community for the help.

☑ This topic is SOLVED.

Need further help from the community? Please ask a new question.

Discussion stats
  • 3 replies
  • 363 views
  • 0 likes
  • 2 in conversation