learningCurve: Fit learning curve for accuracy matrix

In SydneyBioX/scClassify: scClassify: single-cell Hierarchical Classification

Description

Fit learning curve for accuracy matrix

Usage

learningCurve(
  accMat,
  n,
  auto_initial = TRUE,
  a = NULL,
  b = NULL,
  c = NULL,
  d_list = NULL,
  fitmodel = c("nls", "nls_mix", "gam"),
  plot = TRUE,
  verbose = TRUE
)

Arguments

accMat	Matrix of accuracy rate where column indicate different sample size
n	Vector indicates the sample size
auto_initial	whether automatical intialise
a	input the parameter a starting point
b	input the parameter a starting point
c	input the parameter a starting point
d_list	range of d
fitmodel	"nls", "nls_mix", "gam"
plot	indicates whether plot or not
verbose	indicates whether verbose or not

Value

list of results

Author(s)

Yingxin Lin

Examples

set.seed(2019)
n <- seq(20, 10000, 100)
accMat <- do.call(cbind, lapply(1:length(n), function(i){
tmp_n <- rep(n[i], 50)
y <- -2/(tmp_n^0.8) + 0.95 + rnorm(length(tmp_n), 0, 0.02)
}))
res <- learningCurve(accMat = accMat, n)
N <- getN(res, acc = 0.9)

SydneyBioX/scClassify documentation built on Oct. 22, 2021, 4:03 p.m.