Programming the statistical procedures from SAS

Obtaining logit equation for implementing purposes

Accepted Solution Solved
Reply
Contributor
Posts: 36
Accepted Solution

Obtaining logit equation for implementing purposes

Hi

I was wondering if you could help me to obtain the ecuation from a logit regression model, to obtain the individual estimates by hand, since i already have them on a SAS table. The purpose of it, is to give the equation to TI.

  • My ecuation has, continuous, and categorical variables (Trouble is in the last one)
  • Dependent variable is y=1 or y=0
  • example: y = b0 + (b1Married + b2Divorced) + (b3HighSchool + b4College)

                                              Single as "Base"               No Studies as "Base"

For individuals with "Nonbase" characteristics such as Married, Divorced, High School or College, the procedure by hand gives me the estimated "y".

When implementing to production, i find trouble when i have individuals with caracteristics taken as "Base", since the intercept enbody the whole "Base" caracteristics, can´t isolate them and therefore, obtain a different estimated "y".

Could anyone please tell me what to do when having an individual with one or two base caracteristics? How can i get the adjusted "y" i get automatically from enterprise Guide?

Thanks in advanced.

Cesar


Accepted Solutions
Solution
‎12-17-2013 11:03 AM
Grand Advisor
Posts: 16,850

Re: Obtaining logit equation for implementing purposes

When you have "BASE" (usually known as Reference Level) then you only have the intercept and possibly another term. This would be equivalent to what EG would display.

For example for a person who is single with High School education

Y=B0+B3

Person Married with no studies:

Y=B0+B1

Single, No studies

Y=B0

If you're not getting the numbers you expect its likely you've made a mistake either in computing your regression (failing to specify class values with PARAM=REF), setting your output correctly (Y=0 is default regression)

or not doing the correct calculation, not implementing the formula correctly.

The following two links may help you:

Logistic Regression – Quick Tips | Statistics and other stuff from a geek

View solution in original post


All Replies
Solution
‎12-17-2013 11:03 AM
Grand Advisor
Posts: 16,850

Re: Obtaining logit equation for implementing purposes

When you have "BASE" (usually known as Reference Level) then you only have the intercept and possibly another term. This would be equivalent to what EG would display.

For example for a person who is single with High School education

Y=B0+B3

Person Married with no studies:

Y=B0+B1

Single, No studies

Y=B0

If you're not getting the numbers you expect its likely you've made a mistake either in computing your regression (failing to specify class values with PARAM=REF), setting your output correctly (Y=0 is default regression)

or not doing the correct calculation, not implementing the formula correctly.

The following two links may help you:

Logistic Regression – Quick Tips | Statistics and other stuff from a geek

Contributor
Posts: 36

Re: Obtaining logit equation for implementing purposes

Thanks a lot, it worked! This was holding us back.

Cesar

☑ This topic is SOLVED.

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

Discussion stats
  • 2 replies
  • 152 views
  • 0 likes
  • 2 in conversation