Fitter Functions for Linear Models (original) (raw)

lm.fit {stats} R Documentation

Description

These are the basic computing engines called by [lm](../../stats/help/lm.html) used to fit linear models. As they work directly with numeric matrices, they may be more efficient, notably in the case of performing many similar regressions or when inference is not of interest.

[lm](../../stats/help/lm.html) calls lm.fit() as a helper function.

.lm.fit() is a thin wrapper to the "innermost" C code performing the QR decomposition without much checking and should hence be used with care. The same C code is called by lm.fit() and also by[glm.fit](../../stats/help/glm.fit.html)() and [lsfit](../../stats/help/lsfit.html)().

Usage

lm.fit (x, y,    offset = NULL, method = "qr", tol = 1e-7,
                  singular.ok = TRUE, ...)

lm.wfit(x, y, w, offset = NULL, method = "qr", tol = 1e-7,
        wtol = 0, singular.ok = TRUE, ...)

.lm.fit(x, y, tol = 1e-7)

Arguments

x design matrix of dimension n * p.
y vector of observations of length n, or a matrix withn rows.
w vector of weights (length n) to be used in the fitting process for the wfit functions. Weighted least squares is used with weights w, i.e., sum(w * e^2) is minimized.
offset (numeric of length n). This can be used to specify an a priori known component to be included in the linear predictor during fitting.
method currently, only method = "qr" is supported.
tol tolerance for the qr decomposition. Default is 1e-7.
wtol tolerance for treating small positive weights as zero. Default is 0, indicating that only exact zeros are treated as zero.
singular.ok logical. If FALSE, a singular model is an error.
... currently disregarded.

Details

If y is a matrix, offset can be a numeric matrix of the same dimensions, in which case each column is applied to the corresponding column of y.

Value

a [list](../../base/html/list.html) with components (for lm.fit and lm.wfit)

coefficients p vector
residuals n vector or matrix
fitted.values n vector or matrix
effects n vector of orthogonal single-dfeffects. The first rank of them correspond to non-aliased coefficients, and are named accordingly.
weights n vector — only for the *wfit*functions.
rank integer, giving the rank
df.residual degrees of freedom of residuals
qr the QR decomposition, see qr.

Fits without any columns or non-zero weights do not have theeffects and qr components.

.lm.fit() returns a subset of the above, the qr part unwrapped, plus a logical component pivoted indicating if the underlying QR algorithm did pivot.

See Also

[lm](../../stats/help/lm.html) which you should use for linear least squares regression, unless you know better.

Examples

require(utils)
set.seed(129)

n <- 7 ; p <- 2
X <- matrix(rnorm(n * p), n, p) # no intercept!
y <- rnorm(n)
w <- rnorm(n)^2

str(lmw <- lm.wfit(x = X, y = y, w = w))

str(lm. <- lm.fit (x = X, y = y))

## fits w/o intercept:
all.equal(unname(coef(lm(y ~ X-1))),
          unname(coef( lm.fit(X,y))))
all.equal(unname(coef( lm.fit(X,y))),
                 coef(.lm.fit(X,y)))

if(require("microbenchmark")) {
  mb <- microbenchmark(lm(y~X-1), lm.fit(X,y), .lm.fit(X,y))
  print(mb)
  boxplot(mb, notch=TRUE)
}

[Package _stats_ version 4.6.0 Index]