Learning Attribute Patterns in High-Dimensional Structured Latent Attribute Models
This work provides a method for selecting attribute patterns in SLAMs, which is important for researchers in social and biological sciences, but it appears incremental as it builds on existing SLAM frameworks with specific theoretical and methodological enhancements.
The paper tackles the problem of learning significant attribute patterns in high-dimensional structured latent attribute models (SLAMs), addressing theoretical identifiability and proposing a penalized likelihood method with proven selection consistency, as demonstrated through simulation studies and real datasets in educational assessment.
Structured latent attribute models (SLAMs) are a special family of discrete latent variable models widely used in social and biological sciences. This paper considers the problem of learning significant attribute patterns from a SLAM with potentially high-dimensional configurations of the latent attributes. We address the theoretical identifiability issue, propose a penalized likelihood method for the selection of the attribute patterns, and further establish the selection consistency in such an overfitted SLAM with diverging number of latent patterns. The good performance of the proposed methodology is illustrated by simulation studies and two real datasets in educational assessment.