Selection of latent variables for multiple mixed-outcome models

Area of Science:

• Statistics
• Statistical Modeling
• Data Analysis

Background:

• Latent variable models are crucial for analyzing multiple outcomes.
• Model misspecification in latent variable analysis can distort dependence structures and lead to unreliable inference.
• Handling multiple outcomes with varying data types poses significant analytical challenges.

Purpose of the Study:

• To present a general class of latent variable models capable of handling mixed-type outcomes.
• To propose a novel method for simultaneous latent variable selection and parameter estimation.
• To address the limitations of existing models in accurately capturing complex dependence structures.

Main Methods:

• Developed a class of general latent variable models for mixed-type data.
• Proposed a novel simultaneous selection and estimation approach.
• Utilized theoretical analysis to demonstrate estimator properties.

Main Results:

• The proposed estimator is proven to be consistent and asymptotically normal.
• The estimator possesses the oracle property, indicating optimal performance.
• Simulations and a real-world data application confirm the method's practical utility.

Conclusions:

• The novel latent variable modeling approach effectively accommodates mixed-type outcomes.
• The simultaneous selection and estimation method provides reliable and statistically sound inference.
• The approach is validated for practical application in complex data analysis scenarios.