get.cat.xy: Convert original observations to emprical cdf

In qunhualilab/gIDR: Generalized irreproducibile discovery rate

Description

Convert original observations to emprical cdf

Usage

get.cat.xy(x.ori, y.ori, n.missing, miss.sym = 0, top.missing = TRUE,
  multi.fac = 1e-04)

Arguments

x.ori	replicate 1 : a vector of values from original observation
y.ori	replicate 2 : paired with x.ori, a repeated measure having same length with x.ori.
n.missing	the number of observations that miss both x and y.
miss.sym	the numerical symbol to represent a missing observation. It needs to be smaller than the smallest observed value.
top.missing	if missing value is present, top.missing=T; if not consider missing values, top.missing=F.
multi.fac	the cdf on the left side of the lower boundary of the bins

Details

Convert (x, y) from original observations to emprical cdf and unique pairs.

Value

• cat.xy: a data.frame with nrow = number of unique combinations of (x, y). It has the following columns:

  x, y: original values, only unique combinations to define a category.

  factor.x, factor.y: level of factors, only unique values.

  x.cdf, y.cdf: empirical cdf for each category, this is the upper boundary.

  x.cdf.lo, y.cdf.lo: empirical cdf for each category, lower boundary. (i.e. previous upper boundary). To avoid instability at boundary, the lowest x.cdf.lo (y.cdf.lo) is 1/(n.xy+1) and the highest x.cdf (y.cdf) is n.xy/(n.xy+1).

  m: number of observations in each category.

• level.factor: can be used for mapping category format back to original e.g. x.cat[level.factor[1]] == x[1], y.cat[level.factor[1]]==y[1].

Examples

data(chip_seq)
x = chip_seq[,1]
y = chip_seq[,2]
get.cat.xy(x, y, 0, miss.sym=0, top.missing=TRUE, multi.fac=0.0001)

qunhualilab/gIDR documentation built on May 14, 2019, 10:38 a.m.