A1159
Title: High-dimensional multi-study multi-modality covariate-augmented generalized factor model
Authors: Qingzhi Zhong - Jinan University (China) [presenting]
Abstract: Latent factor models that integrate data from multiple studies or modalities have garnered considerable attention across various disciplines. However, existing methods predominantly focus either on multi-study integration or multi-modality integration, rendering them insufficient for analyzing diverse modalities measured across multiple studies. To address this limitation and cater to practical needs, a high-dimensional generalized factor model is introduced that seamlessly integrates multi-modality data from multiple studies while accommodating additional covariates. A thorough investigation of the identifiability conditions is conducted to enhance model interpretability. To tackle the complexity of high-dimensional nonlinear integration caused by four large latent random matrices, a variational lower bound is utilized to approximate the observed log-likelihood by employing a variational posterior distribution. By profiling the variational parameters, asymptotical properties of estimators for model parameters are established using M-estimation theory. Furthermore, a computationally efficient variational EM algorithm is devised to execute the estimation process and a criterion to determine the optimal number of both study-shared and study-specific factors. Extensive simulation studies and a real-world application demonstrate that the proposed method significantly outperforms existing methods in terms of estimation accuracy and computational efficiency.
