Thread Subject: Function for OLS regression with a dummy variable

Dear All,

I want to estimate a simple linear regression (get parameter and error estimates) that includes a dummy variable (say Dv) taking values of 0 and 1, of the following form:

y=constant+a*x+f*Dv+errors

Note that the variable Dv is predetermined depending on the time that a break occurs (ToB)
e.g. Dv(1:ToB,1)=0;
Dv(ToB:length(x),1)=1;

where say ToB=100 and length(x)=800


The question is which function should I use to get the OLS parameter estimates (constant, s,f,errors)?

Thank you very much for your time and help!

"Chris" <AzizC0rn@bluestring.com> wrote in message <hc1bep$1d5$1@fred.mathworks.com>...
> Dear All,
>
> I want to estimate a simple linear regression (get parameter and error estimates) that includes a dummy variable (say Dv) taking values of 0 and 1, of the following form:
>
> y=constant+a*x+f*Dv+errors
>
> Note that the variable Dv is predetermined depending on the time that a break occurs (ToB)
> e.g. Dv(1:ToB,1)=0;
> Dv(ToB:length(x),1)=1;
>
> where say ToB=100 and length(x)=800
>
>
> The question is which function should I use to get the OLS parameter estimates (constant, s,f,errors)?
>
> Thank you very much for your time and help!

I think I found it (thought that X must be a vector):
X=[x Dv];
[b,bint,r,rint] = regress(y,X)