|
Activity Number:
|
496
- Dimension Reduction
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Statistical Learning and Data Science
|
|
Abstract #313217
|
|
|
Title:
|
Learning Hierarchical Structures in Latent Attribute Models
|
|
Author(s):
|
Chenchen Ma* and Gongjun Xu
|
|
Companies:
|
University of Michigan and University of Michigan
|
|
Keywords:
|
latent variable models;
mixture models;
hierarchical structures;
regularization;
sparse estimation
|
|
Abstract:
|
Hierarchical Latent Attribute Models (HLAMs) are a special family of restricted discrete latent variable models widely used in social and biomedical sciences. In many applications, certain hierarchical constraints are put on allowable patterns of the latent attributes. For instance, some lower-level attributes are assumed to be prerequisites for higher-level attributes. This paper considers the problem of learning latent hierarchical structures from noisy observations with minimal model assumptions. A regularized latent attribute model is proposed, and an EM-type algorithm is developed for efficient and scalable computation. We further show that the proposed approach enjoys nice theoretical properties. The good performance of the proposed methodology is illustrated by extensive simulation studies and a real dataset in educational assessment.
|
Authors who are presenting talks have a * after their name.
