Efficient Metropolis-Hastings Robbins-Monro Algorithm for High-Dimensional Diagnostic Classification Models

Area of Science:

• Psychometrics
• Computational Statistics
• Educational Measurement

Background:

• The Expectation-Maximization (EM) algorithm is standard for parameter estimation in diagnostic classification models (DCMs).
• EM's computational complexity, O(2^K), becomes prohibitive for models with many attributes (large K).

Purpose of the Study:

• To introduce a computationally efficient algorithm for parameter estimation in high-dimensional DCMs.
• To compare the performance of the proposed algorithm against existing methods.

Main Methods:

• Developed an efficient Metropolis-Hastings Robbins-Monro (eMHRM) algorithm with O(K+1) complexity.
• Utilized the Robbins-Monro algorithm for approximating item and structural parameters, avoiding nonlinear optimization.
• Conducted simulation studies to compare eMHRM with EM and standard Metropolis-Hastings (MH).

Main Results:

• The eMHRM algorithm demonstrated significantly greater computational efficiency than both EM and MH algorithms.
• eMHRM provided improved parameter estimates compared to EM, particularly for models with a large number of attributes (large K).

Conclusions:

• The eMHRM algorithm is a promising and efficient method for parameter estimation in high-dimensional DCMs.
• This approach addresses the computational limitations of traditional methods for complex diagnostic models.