Std Error for a coefficient
In linear regression the standard error for a coefficient is a function of the errors, variance, and the sample size. The formula (see links below for an easier to read version) is:
s = sqrt[ ∑ ((e_i)^2/(n-2)) / (∑ (x_i - xbar)^2) ]
where
x_i = the ith observation
e_i = the ith error = y_i - yhat_i
n = sample size
xbar = mean of the x values
s = sqrt[ ∑ ((e_i)^2/(n-2)) / (∑ (x_i - xbar)^2) ]
where
x_i = the ith observation
e_i = the ith error = y_i - yhat_i
n = sample size
xbar = mean of the x values
Source:
posted by jerdavis at 9:26 AM
0 Comments:
Post a Comment
<< Home