estip2: Unidimensional Item Response Theory parameter estimation
In takuizum/irtfun2: Some useful functions for IRT

Description Usage Arguments Value Examples

Unidimensional Item Response Theory parameter estimation

estip2(
  x,
  fc = 3,
  IDc = 1,
  Gc = NULL,
  bg = 1,
  Ntheta = 31,
  D = 1.702,
  method = "Fisher_Scoring",
  model = "2PL",
  max_func = "N",
  rm_list = NULL,
  mu_th = 0,
  sigma_th = 1,
  min_th = -4,
  max_th = 4,
  eEM = 0.001,
  eMLL = 0.001,
  maxiter_em = 100,
  th_dist = "normal",
  fix_a = 1,
  mu_a = 0,
  sigma_a = 1,
  mu_b = 0,
  sigma_b = 1,
  mu_c = 0,
  sigma_c = 1,
  w_c = 1,
  alpha = 0.5,
  lambda = 1,
  print = 0
)

`x`	DataFrame.
`fc`	the first column.
`IDc`	the ID column.
`Gc`	the grade column.
`bg`	a mumber of base grade.
`Ntheta`	the number of the nodes of theta dist.
`D`	a scale constant.
`method`	the method of optimiser. Default is "Fisher_Scoring", but `optim` function also be able to use.
`model`	a model vector
`max_func`	a character of object function. "N" is MML-EM, "B" is marginal Bayes and "R" is Regularized MML.
`rm_list`	a vector of item U want to remove for estimation. NOT list.
`mu_th`	a hyper parameter of normal dist for theta
`sigma_th`	a hyper parameter of normal dist for theta
`min_th`	a minimum value of theta distribution
`max_th`	a maximum value of theta distribution
`eEM`	a convergence criterion related to item parameters in EM cycle.
`eMLL`	a convergence criterion related to negative twice log likelihood in EM cycle.
`maxiter_em`	the number of iteration of EM cycle.
`th_dist`	a type of theta dist."normal" or "empirical"
`fix_a`	a fix parameter for slope parameter of 1PLM
`mu_a`	a hyper parameter for slope parameter prior distribution(lognormal) in marginal Bayes estimation.
`sigma_a`	a hyper parameter for slope parameter prior distribution(lognormal) in marginal Bayes estimation.
`mu_b`	a hyper parameter for location parameter prior distribution(normal) in marginal Bayes estimation.
`sigma_b`	a hyper parameter for location parameter prior distribution(normal) in marginal Bayes estimation.
`mu_c`	a hyper parameter for lower asymptote parameter prior distribution(beta) in marginal Bayes estimation.
`sigma_c`	a hyper parameter for lower asymptote parameter prior distribution(beta) in marginal Bayes estimation.
`w_c`	a weight parameter for lower asymptote parameter prior distribution(beta) in marginal Bayes estimation.
`alpha`	tuning parameter of elastic net penalty.
`lambda`	tuning parameter of elastic net penalty.
`print`	How much information you want to display? from 1 to 3. The larger, more information is displayed.

the output is a list that has item parameters data.frame and these Standard Error.

# MMLE
res1 <- estip2(x=sim_data_2, Gc=NULL, fc=2, Ntheta=21)
# check the parameters
res1$para
res1$SE

# Multigroup MMLE
res2 <- estip2(x=sim_dat_st, Gc=2, bg=3, fc=3, Ntheta=10)

# Marginal Bayes
res3 <- estip2(x=sim_data_2, Gc=NULL, fc=2, Ntheta=21, max_func="B")

# Regularized MMLE
res4 <- estip2(x=sim_data_2, Gc=NULL, fc=2, Ntheta=21, max_func="R")