Package: ML2Pvae 1.0.0.1

ML2Pvae: Variational Autoencoder Models for IRT Parameter Estimation

Based on the work of Curi, Converse, Hajewski, and Oliveira (2019) <doi:10.1109/IJCNN.2019.8852333>. This package provides easy-to-use functions which create a variational autoencoder (VAE) to be used for parameter estimation in Item Response Theory (IRT) - namely the Multidimensional Logistic 2-Parameter (ML2P) model. To use a neural network as such, nontrivial modifications to the architecture must be made, such as restricting the nonzero weights in the decoder according to some binary matrix Q. The functions in this package allow for straight-forward construction, training, and evaluation so that minimal knowledge of 'tensorflow' or 'keras' is required.

Authors:Geoffrey Converse [aut, cre, cph], Suely Oliveira [ctb, ths], Mariana Curi [ctb]

ML2Pvae_1.0.0.1.tar.gz
ML2Pvae_1.0.0.1.zip(r-4.7)ML2Pvae_1.0.0.1.zip(r-4.6)ML2Pvae_1.0.0.1.zip(r-4.5)
ML2Pvae_1.0.0.1.tgz(r-4.6-any)ML2Pvae_1.0.0.1.tgz(r-4.5-any)
ML2Pvae_1.0.0.1.tar.gz(r-4.7-any)ML2Pvae_1.0.0.1.tar.gz(r-4.6-any)
ML2Pvae_1.0.0.1.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION
card.svg |card.png
ML2Pvae/json (API)

# Install 'ML2Pvae' in R:

install.packages('ML2Pvae', repos = c('https://converseg.r-universe.dev', 'https://cloud.r-project.org'))

Datasets:

correlation_matrix - Simulated latent abilities correlation matrix
diff_true - Simulated difficulty parameters
disc_true - Simulated discrimination parameters
q_matrix - Simulated Q-matrix
responses - Response data
theta_true - Simulated ability parameters

On CRAN:

This package does not link to any Github/Gitlab/R-forge repository. No issue tracker or development information is available.

2.00 score 4 scripts 265 downloads 5 exports 34 dependencies

Last updated from:692b9d583c. Checks:7 NOTE, 2 OK. Indexed: yes.

Target	Result	Time
linux-devel-x86_64	NOTE	124
source / vignettes	OK	196
linux-release-x86_64	NOTE	129
macos-release-arm64	NOTE	117
macos-oldrel-arm64	NOTE	130
windows-devel	NOTE	85
windows-release	NOTE	120
windows-oldrel	NOTE	65
wasm-release	OK	105

Exports:build_vae_correlated build_vae_independent get_ability_parameter_estimates get_item_parameter_estimates train_model

Dependencies:backports base64enc cli config generics glue here jsonlite keras lattice lifecycle magrittr Matrix png processx ps R6 rappdirs Rcpp RcppTOML reticulate rlang rprojroot rstudioapi tensorflow tfautograph tfprobability tfruns tidyselect vctrs whisker withr yaml zeallot

ML2Pvae: Variational Autoencoder Models for IRT Parameter Estimation

Rendered fromml2p_vae_vignette.pdf.asisusingR.rsp::asis

Last update: 2020-11-16
Started: 2020-11-16

Help page	Topics
Display a message upon loading package	.onLoad
Build the encoder for a VAE	build_hidden_encoder
Build a VAE that fits to a normal, full covariance N(m,S) latent distribution	build_vae_correlated
Build a VAE that fits to a standard N(0,I) latent distribution with independent latent traits	build_vae_independent
Simulated latent abilities correlation matrix	correlation_matrix
Simulated difficulty parameters	diff_true
Simulated discrimination parameters	disc_true
Feed forward response sets through the encoder, which outputs student ability estimates	get_ability_parameter_estimates
Get trainable variables from the decoder, which serve as item parameter estimates.	get_item_parameter_estimates
ML2Pvae: A package for creating a VAE whose decoder recovers the parameters of the ML2P model. The encoder can be used to predict the latent skills based on assessment scores.	ML2Pvae
A custom kernel constraint function that forces nonzero weights to be equal to one, so the VAE will estimate the 1-parameter logistic model. Nonzero weights are determined by the Q matrix.	q_1pl_constraint
A custom kernel constraint function that restricts weights between the learned distribution and output. Nonzero weights are determined by the Q matrix.	q_constraint
Simulated Q-matrix	q_matrix
Response data	responses
A reparameterization in order to sample from the learned multivariate normal distribution of the VAE	sampling_correlated
A reparameterization in order to sample from the learned standard normal distribution of the VAE	sampling_independent
Simulated ability parameters	theta_true
Trains a VAE or autoencoder model. This acts as a wrapper for keras::fit().	train_model
A custom loss function for a VAE learning a multivariate normal distribution with a full covariance matrix	vae_loss_correlated
A custom loss function for a VAE learning a standard normal distribution	vae_loss_independent
Give error messages for invalid inputs in exported functions.	validate_inputs

Package: ML2Pvae 1.0.0.1

ML2Pvae: Variational Autoencoder Models for IRT Parameter Estimation

Citation

Readme and manuals

Help Manual

Usage by other packages (reverse dependencies)