,
Steven Andrew Culpepper [aut, cph]
| Title: | Sparse Latent Class Model for Cognitive Diagnosis |
|---|---|
| Description: | Perform a Bayesian estimation of the exploratory Sparse Latent Class Model for Binary Data described by Chen, Y., Culpepper, S. A., and Liang, F. (2020) <doi:10.1007/s11336-019-09693-2>. |
| Authors: | James Joseph Balamuta [aut, cre, cph]
,
Steven Andrew Culpepper [aut, cph]
|
| Maintainer: | James Joseph Balamuta <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 0.1.0 |
| Built: | 2024-11-11 03:13:26 UTC |
| Source: | https://github.com/tmsalab/slcm |
Generate attribute pattern table header
attribute_pattern_table_header(k, m = 2, order = k)attribute_pattern_table_header(k, m = 2, order = k)
k |
Number of Attributes. |
m |
Number of Categories. Default 2 or dichotomous response. |
order |
Order of the table. Default |
Return a matrix containing the class table
# K = 3 attribute_pattern_table_header(3) # K = 4 attribute_pattern_table_header(4)# K = 3 attribute_pattern_table_header(3) # K = 4 attribute_pattern_table_header(4)
Custom printing class to reveal features of the fitted SLCM.
## S3 method for class 'slcm' print(x, digits = max(3L, getOption("digits") - 3L), ...)## S3 method for class 'slcm' print(x, digits = max(3L, getOption("digits") - 3L), ...)
x |
the |
digits |
the number of significant digits |
... |
further arguments passed to or from other methods. |
Print details and estimates found within the fitted SLCM.
Return the model invisibly (via invisible())
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>
slcm( y, k, burnin = 1000L, chain_length = 10000L, psi_invj = c(1, rep(2, 2^k - 1)), m0 = 0, bq = 1 )slcm( y, k, burnin = 1000L, chain_length = 10000L, psi_invj = c(1, rep(2, 2^k - 1)), m0 = 0, bq = 1 )
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. |
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.
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.
# Use a demo data set from the paper 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) print(model_reasoning)# Use a demo data set from the paper 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) print(model_reasoning)