Construct a lazy residual matrix

Returns a DelayedArray-backed matrix representing Y - invLink(tcrossprod(Beta, design)) without materializing the full result.

Usage

ResidualMatrixGLM(
  Y,
  X,
  Beta,
  family = "gaussian",
  nb.theta = NA_real_,
  fit,
  type = c("deviance", "pearson", "response", "response.se"),
  chunkSize = 1000
)

# S4 method for class 'DelayedMatrix,matrix,matrix,ANY,ANY,ANY'
ResidualMatrixGLM(
  Y,
  X,
  Beta,
  family = "gaussian",
  nb.theta = NA_real_,
  type = c("deviance", "pearson", "response", "response.se"),
  chunkSize = 1000
)

# S4 method for class 'ANY,matrix,matrix,ANY,ANY,ANY'
ResidualMatrixGLM(
  Y,
  X,
  Beta,
  family = "gaussian",
  nb.theta = NA_real_,
  type = c("deviance", "pearson", "response", "response.se"),
  chunkSize = 1000
)

# S4 method for class 'DelayedMatrix,matrix,missing,missing,missing,modelFitsResponses'
ResidualMatrixGLM(
  Y,
  X,
  Beta,
  family = "gaussian",
  nb.theta = NA_real_,
  fit,
  type = c("deviance", "pearson", "response", "response.se"),
  chunkSize = 1000
)

# S4 method for class 'ANY,matrix,missing,missing,missing,modelFitsResponses'
ResidualMatrixGLM(
  Y,
  X,
  Beta,
  family = "gaussian",
  nb.theta = NA_real_,
  fit,
  type = c("deviance", "pearson", "response", "response.se"),
  chunkSize = 1000
)

Arguments

Y

A matrix-like object coercible to DelayedArray.

X

A matrix-like object coercible to DelayedArray.

Beta

A numeric matrix or object coercible to matrix.

family

GLM family string

nb.theta

array of overdispersion parameters for negative binomial regressions

fit

multivariate regression fit of class modelFitsResponses from glmFitResponses(), or similar

type

type of residual: c("deviance", "pearson", "response", "response.se")

chunkSize

number of rows to read into memory at a time

Value

A DelayedMatrix.

Details

Requested blocks are computed on demand through the seed's extract_array() method.

See also

residuals.glm(), computeResiduals()

Examples

library(BatchRegression)
library(fastglmm)
#> Loading required package: nlme

n <- 100
m <- 5
nc <- 2
set.seed(1)
Y <- matrix(rpois(n * m, 100), m, n)
X <- matrix(rnorm(n * nc), n, nc)
X <- cbind(1, X) # intercept term
colnames(X) <- seq(ncol(X))
rownames(Y) <- seq(m)

# negative.binomial
fam <- negative.binomial(4)
fit <- glmFitResponses(Y, X, fam)

# residuals
R <- ResidualMatrixGLM(Y, X, fit = fit)