How to use lm.fit instead on lm [closed]

问题

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.

Closed 6 years ago.

I want to use lm.fit for speed, but second version gives NAs

 sum <-   summary(lm(y~x))
 slope <-  sum$coefficients[2]

or

 sum <-   lm.fit(as.matrix(x,ncol=1),y)
 slope <-  sum$coefficients[2]

EDIT 1

I now see that sum$coefficients only has 1 value. Why is that and what is it? The help doesn't explain this

回答1:

It would be easier to help with a reproducible example.

However, my guess is that you are missing an intercept in the second case. Try lm.fit(cbind(1,x),y) and see if that gives you the comparison that you are looking for.

Other differences could be due to the other preprocessing that lm does before calling lm.fit, but we don't know what that is without seeing x. Things like removing missing values, expanding a factor into dummy variables could also result in differences.

来源：https://stackoverflow.com/questions/16675614/how-to-use-lm-fit-instead-on-lm