how to generate a new equation by using regression coefficient

[WEN-CHIA HUANG]

Hi, I am wondering if I can generate a new equation y = a + b1 * gender + b2 * age where the "a", "b1", "b2" is from the regression coefficient of data1.
Then, I need to use new observations from data2 to put into the new equation to get new y.

I already run the regression results of data1 and store the coefficient into matrix by the following command:
local b1 = m[1,2]
local b2 = m[1,3]
However, I do not know how to generate a new equation by using regression coefficient of data1.
Because I still need to put observations from data2 into this equation, would I need to save the equation into a new dofile?
Does anyone have idea to slove this issue?
Thank you!

[Joseph Coveney]

. . .the "a", "b1", "b2" is from the regression coefficient of data1.
Then, I need to use new observations from data2 to put into the new equation to get new y.

To get new y from the regression coefficients for data1, use predict. Something like the following.

use data1
generate byte data1 = 1

append using data2
replace data1 = 0 if missing(data1)

regress y i.gender c.age if data1
predict double new_y, xb // <= here

list data1 gender age y new_y, noobs sepby(data1)

[WEN-CHIA HUANG]

To get new y from the regression coefficients for data1, use predict. Something like the following.

use data1
generate byte data1 = 1

append using data2
replace data1 = 0 if missing(data1)

regress y i.gender c.age if data1
predict double new_y, xb // <= here

list data1 gender age y new_y, noobs sepby(data1)

Thank you very much! It helps me a lot.

[Joseph Coveney]

You're welcome.

One thing that you might not have noticed because I was emphasizing the predict line of code, but the dataset-indicating variable is used primarily to restrict the regression model to the data1 dataset. I neglected to emphasize that above, and so I'll point it out here (see below),

regress y i.gender c.age if data1

[WEN-CHIA HUANG]

You're welcome.

One thing that you might not have noticed because I was emphasizing the predict line of code, but the dataset-indicating variable is used primarily to restrict the regression model to the data1 dataset. I neglected to emphasize that above, and so I'll point it out here (see below),

regress y i.gender c.age if data1

Thanks! I've revised my code and successfully run the regression.
I did notice " if data1 " in your code, and I also did it in mine.
Thanks for the reminder.