sety: Set y-Values

In lm.br: Linear Model with Breakpoint

Set y-Values

Description

Reset the response values in the C++ object.

Usage

## S4 method for signature 'Cpp_Clmbr'
sety( rWy )

Arguments

rWy	vector of 'y' values, pre-multiplied by the square-root of 'weights'.

Details

The 'rWy' vector is simply the y-vector if the model does not specify weights. The square-root of a vector 'W' is the vector 'rW' of the square-roots of the elements of 'W'. The square-root of a matrix 'W' here is the matrix 'rW' such that rW*rW = W (a stricter definition than rW*transpose(rW) = W).

Note

The pre-multiplied vector is more convenient as input during simulation tests. 'sety' changes the y-values only for the accessor functions 'sl', 'ci', 'cr' and 'mle'. 'rW' is the inverse square-root if 'inverse' was TRUE in the 'lm.br' call.

Examples

#  A simulation test
x <- c( 1.0, 1.1, 1.3, 1.7, 2.4, 3.9, 5.7, 7.6, 8.4, 8.6 )
y <- x
LLmodel <- lm.br( y ~ x )    
countCLR <- countAF <- 0
theta <- 3
for( i in 1:10000 )  {
  y <- 0 + (-1.)*pmin(x-theta,0) + (0.5)*pmax(x-theta,0) + rnorm(10)
  LLmodel$sety( y )
  stest <- LLmodel$sl( theta, 'clr', .0001, "V" )
  if( stest > 0.05 )  countCLR <- countCLR + 1
  stest <- LLmodel$sl( theta, 'af', .0001, "V" )
  if( stest > 0.05 )  countAF <- countAF + 1
  if( floor(i/1000) - i/1000 == 0 ) cat(i, countCLR/i, countAF/i, "\n")
}

lm.br documentation built on March 19, 2024, 3:08 a.m.