Least squares regression sum must be zero

chengeto

New member
Joined
Feb 28, 2009
Messages
49
On an old exam paper they have this proof question. How do l work it out ?
 
The question says show that the sum of residuals in least squares regression must always be zero ?
 
chengeto said:
The question says show that the sum of residuals in least squares regression must always be zero ?

Please share with us your work/thoughts, indicating exactly where you are stuck - so that we know where to begin to help you.
 
Subhotosh Khan said:
chengeto said:
The question says show that the sum of residuals in least squares regression must always be zero ?

Please share with us your work/thoughts, indicating exactly where you are stuck - so that we know where to begin to help you.


Here is my attempt:

=\(\displaystyle y_{i}=a+bx_{i}+\epsilon\)

=\(\displaystyle \epsilon=y_{i}-a-bx_{i}\)

=\(\displaystyle \epsilon^2_{1}+\epsilon^2_{2}+\dots+\epsilon^2_{n}\)

=\(\displaystyle (y_{1}-a-bx_{1})^2+ (y_{2}-a-bx_{2})^2+\dots(y_{n}-a-bx_{n})^2\)


It is here that l get stuck. I don't know what to do ?
 
Top