Estimating the DINA model parameters using the No-U-Turn Sampler.

摘要: The deterministic inputs, noisy, "and" gate (DINA) model is a popular cognitive diagnosis (CDM) in psychology and psychometrics used to identify test takers' profiles with respect set of latent attributes or skills. In this work, we propose an estimation method for the DINA No-U-Turn Sampler (NUTS) algorithm, extension Hamiltonian Monte Carlo (HMC) method. We conduct simulation study order evaluate parameter recovery efficiency new Markov chain compare it two other Bayesian methods, Metropolis Hastings Gibbs sampling algorithms, frequentist method, using Expectation-Maximization (EM) algorithm. results indicated that NUTS algorithm employed properly recovers all parameters accurate simulated scenarios. apply methodology mental health area develop classification respondents Beck Depression Inventory. implementation applied psychological tests has potential improve medical diagnostic process.