Abstract:The problem of causal discovery is especially challenging when the variables of interest cannot be directly measured. In measurement models, the measured variables were generated by latent causal variables that are causally related to each other, and by estimating the measurement model from measured data, one is able to recover the latent variables and their causal relations. In this paper, we provide precise sufficient identifiability conditions for the linear pure measurement model, and show what information of the causal structure can be recovered from observed data without prior knowledge of data distributions. In particular, we first show that, based on second-order statistics, although the pure measurement model is in general not fully identifiable under the assumption of two pure measurement variables, we can identify the set of all candidate measurement models. We then further prove that the pure measurement model can be identified uniquely based on higher-order statistics. Next, to address more general situations, we offer the identifiability conditions of linear measurement models with arbitrary noise distributions. We finally develop a unified method to learn pure measurement models from data. Experimental results on both synthetic and real-world data demonstrate the usefulness of our theory and the effectiveness of our approach.

Identification of the linear factor model

Identifiability of latent-variable and structural-equation models: from linear to nonlinear

IDENTIFIABILITY OF BIFACTOR MODELS

Causal Discovery of 1-Factor Measurement Models in Linear Latent Variable Models with Arbitrary Noise Distributions.

Weak Identification in Low-Dimensional Factor Models with One or Two Factors

Bounded support in linear random coefficient models: Identification and variable selection

Optimal Estimation of Large-Dimensional Nonlinear Factor Models

Non-parametric Regression Among Factor Scores: Motivation and Diagnostics for Nonlinear Structural Equation Models

Structured Latent Factor Analysis for Large-scale Data: Identifiability, Estimability, and Their Implications

State-Varying Factor Models of Large Dimensions

Estimation of a Factor-Augmented Linear Model with Applications Using Student Achievement Data

On Time-Varying Factor Models: Estimation and Testing

Identifying Observed Factors in Approximate Factor Models: Estimation and Hypothesis Testing

Theoretical Analysis and Comparison of Several Criteria on Linear Model Dimension Reduction

A Framework for Eliciting, Incorporating, and Disciplining Identification Beliefs in Linear Models

On the Complexity of Identification in Linear Structural Causal Models

Covariance Function Estimation for High-Dimensional Functional Time Series with Dual Factor Structures

Sparsity Identification for High-Dimensional Partially Linear Model with Measurement Error.

Structured prior distributions for the covariance matrix in latent factor models