slcm {slcm}

Sparse Latent Class Model for Cognitive Diagnosis (SLCM)

Description

Performs the Gibbs sampling routine for a sparse latent class model as described in Chen et al. (2020) <doi: 10.1007/s11336-019-09693-2>

Usage

slcm(
  y,
  k,
  burnin = 1000L,
  chain_length = 10000L,
  psi_invj = c(1, rep(2, 2^k - 1)),
  m0 = 0,
  bq = 1
)

Arguments

y

Item Matrix

k

Dimension to estimate for Q matrix

burnin

Amount of Draws to Burn

chain_length

Number of Iterations for chain.

psi_invj, m0, bq

Additional tuning parameters.

Details

The estimates list contains the mean information from the sampling procedure. Meanwhile, the chain list contains full MCMC values. Lastly, the details list provides information regarding the estimation call.

Value

An slcm object containing three named lists:

• estimates

  • beta: Average beta coefficients

  • theta: Average theta coefficients

  • delta: Average activeness of coefficients

  • class: Average class membership

  • pi: Average attribute class probability.

  • omega: Average omega

  • q: Average activeness of Q matrix entries based on heuristic transformation.

  • m2ll: Average negative two times log-likelihood

• chain

  • theta: theta coefficients iterations

  • beta: beta coefficients iterations

  • class: class membership iterations

  • pi: attribute class probability iterations

  • omega: omega iterations

  • m2ll: Negative two times log-likelihood iterations

• details

  • n: Number of Subjects

  • j: Number of Items

  • k: Number of Traits

  • l1: Slab parameter

  • m0, bq: Additional tuning parameters

  • burnin: Number of Iterations to discard

  • chain_length: Number of Iterations to keep

  • runtime: Duration of model run inside of the C++ code. (Does not include summarization of MCMC chain.)

  • package_version: Version of the package the SLCM model was fit with.

  • date_time: Date and Time the model was fit.

Examples

# Use a demo data set from the paper
if(requireNamespace("edmdata", quietly = TRUE)) {
  data("items_matrix_reasoning", package = "edmdata")
  
  burnin = 50        # Set for demonstration purposes, increase to at least 1,000 in practice.
  chain_length = 100 # Set for demonstration purposes, increase to at least 10,000 in practice.  
  
  model_reasoning = slcm(items_matrix_reasoning, k = 4, 
                         burnin = burnin, chain_length = chain_length)
}