adjBy2ptReg {wrMisc}

Linear rescaling of numeric vertor or matrix

Description

adjBy2ptReg takes data within window defined by 'lims' and determines linear transformation so that these points get the regression characteristics 'regrTo', all other points (ie beyond the limits) will follow the same transformation. In other words, this function performs 'linear rescaling', by adjusting (normalizing) the vector 'dat' by linear regression so that points falling in 'lims' (list with upper & lower boundaries) will end up as 'regrTo'.

Usage

adjBy2ptReg(
  dat,
  lims,
  regrTo = c(0.1, 0.9),
  refLines = NULL,
  silent = FALSE,
  debug = FALSE,
  callFrom = NULL
)

Arguments

dat

numeric vector, matrix or data.frame

lims

(list, length=2) should be list giving limits (list(lo=c(min,max),hi=c(min,max)) in data allowing identifying which points will be used for determining slope & offset

regrTo

(numeric, length=2) to which characteristics data should be regressed

refLines

(NULL or integer) optional subselection of lines of dat (will be used internal as refDat)

silent

(logical) suppress messages

debug

(logical) display additional messages for debugging

callFrom

(character) allow easier tracking of messages produced

Value

This function returns a matrix (of same dimensions as inlut matrix) with normalized values

See Also

normalizeThis

Examples

set.seed(2016); dat1 <- round(runif(50,0,100),1)
## extreme values will be further away :
adjBy2ptReg(dat1,lims=list(c(5,9), c(60,90)))
plot(dat1, adjBy2ptReg(dat1, lims=list(c(5,9),c(60,90))))