Nonstationary multivariate Gaussian processes for electronic health records

Meng, Rui; Soper, Braden; Lee, Herbert K. H.; Vincent, Jean‐Louis; Greene, J; Ray, Priyadip

doi:10.1016/j.jbi.2021.103698

Cited by 12 publications

(6 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This mixing mechanism with independent latent processes is called spatially varying linear model of corregionalization (SVLMC) [14] in spatial statistics literature. Recently, [23] propose a general regression framework based on this mixing mechanism and get a successful implementation of the analysis in electronic health records. In addition, replacing latent processes f (x) with noisy latent processes f (x) + σ f , assuming homogeneous noise such that Σ = σ 2 y I P and modeling each element of W via a Gaussian process lead the SLMM to be the exact Gaussian process regression network (GPRN) in [34].…”

Section: Stochastic Linear Mixing Modelmentioning

confidence: 99%

“…The adaptive mixture of Gaussian processes allows it account for input-dependent correlations between outputs. Recently, [23] leveraged the adaptive linear projection structure and propose a general regression framework to deal with input-dependent correlation, scale and smoothness of outputs. All of those models assume that the mixing coefficients are input-dependent and data live around a Q dimensional linear subspace, where Q < P and P is the output dimension.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bayesian Inference in High-Dimensional Time-Serieswith the Orthogonal Stochastic Linear Mixing Model

Meng¹,

Bouchard²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Many modern time-series datasets contain large numbers of output response variables sampled for prolonged periods of time. For example, in neuroscience, the activities of 100s-1000's of neurons are recorded during behaviors and in response to sensory stimuli. Multi-output Gaussian process models leverage the nonparametric nature of Gaussian processes to capture structure across multiple outputs. However, this class of models typically assumes that the correlations between the output response variables are invariant in the input space. Stochastic linear mixing models (SLMM) assume the mixture coefficients depend on input, making them more flexible and effective to capture complex output dependence. However, currently, the inference for SLMMs is intractable for large datasets, making them inapplicable to several modern time-series problems. In this paper, we propose a new regression framework, the orthogonal stochastic linear mixing model (OSLMM) that introduces an orthogonal constraint amongst the mixing coefficients. This constraint reduces the computational burden of inference while retaining the capability to handle complex output dependence. We provide Markov chain Monte Carlo inference procedures for both SLMM and OSLMM and demonstrate superior model scalability and reduced prediction error of OSLMM compared with state-of-the-art methods on several real-world applications. In neurophysiology recordings, we use the inferred latent functions for compact visualization of population responses to auditory stimuli, and demonstrate superior results compared to a competing method (GPFA). Together, these results demonstrate that OSLMM will be useful for the analysis of diverse, large-scale time-series datasets.Preprint. Under review.

show abstract

Section: Stochastic Linear Mixing Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Bayesian Inference in High-Dimensional Time-Serieswith the Orthogonal Stochastic Linear Mixing Model

Meng¹,

Bouchard²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Various approaches to disease progression modeling have been proposed in the literature. These approaches range from deterministic approaches based on differential equations, 11 statistical approaches such as autoregressive models, 12 hidden Markov models, 13 and Gaussian processes, 14 , 15 deep learning methods such as recurrent neural networks, 16 and computational simulation methods such as discrete event simulations (DESs). 17 , 18 The choice of modeling approach depends on the degree of knowledge of the underlying disease mechanism, the stochasticity and heterogeneity of disease symptoms, the number of samples available for parameter estimation, and the need for model interpretability.…”

Section: Introductionmentioning

confidence: 99%

Dynamic modeling of hospitalized COVID-19 patients reveals disease state–dependent risk factors

Soper

Cadena

Nguyen

et al. 2022

Journal of the American Medical Informatics Association

Self Cite

View full text Add to dashboard Cite

Objective The study sought to investigate the disease state–dependent risk profiles of patient demographics and medical comorbidities associated with adverse outcomes of SARS-CoV-2 coronavirus infections. Materials and Methods A covariate-dependent, continuous-time hidden Markov model with four states (moderate, severe, discharged, and deceased) was used to model the dynamic progression of COVID-19 during the course of hospitalization. All model parameters were estimated using the electronic health records of 1,362 patients from ProMedica Health System admitted between March 20, 2020 and December 29, 2020 with a positive nasopharyngeal PCR test for SARS-CoV-2. Demographic characteristics, comorbidities, vital signs and laboratory test results were retrospectively evaluated to infer a patient’s clinical progression. Results The association between patient-level covariates and risk of progression was found to be disease state dependent. Specifically, while being male, being Black or having a medical comorbidity were all associated with an increased risk of progressing from the moderate disease state to the severe disease state, these same factors were associated with a decreased risk of progressing from the severe disease state to the deceased state. Discussion Recent studies have not included analyses of the temporal progression of COVID-19, making the current study a unique modeling-based approach to understand the dynamics of COVID-19 in hospitalized patients. Conclusion Dynamic risk stratification models have the potential not only to improve clinical outcomes in COVID-19, but also a myriad of other acute and chronic diseases that, to date, have largely been assessed only by static modeling techniques.

show abstract

“…HMM methods can be formulated as continuous-time models, where transition probabilities depend on inter-measurement intervals [36,21]. Gaussian Process (GP) models capture temporal dynamics and have been developed for EHR data [7,1,27,18]. RNN approaches can be used by predicting the time until an event from the current input and the time since the last input [5].…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Probabilistic Models for Sequential Electronic Health Records

Kaplan¹,

Greene²,

Vincent³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

We develop an unsupervised probabilistic model for heterogeneous Electronic Health Record (EHR) data. Utilizing a mixture model formulation, our approach directly models sequences of arbitrary length, such as medications and laboratory results. This allows for subgrouping and incorporation of the dynamics underlying heterogeneous data types. The model consists of a layered set of latent variables that encode underlying structure in the data. These variables represent subject subgroups at the top layer, and unobserved states for sequences in the second layer. We train this model on episodic data from subjects receiving medical care in the Kaiser Permanente Northern California integrated healthcare delivery system. The resulting properties of the trained model generate novel insight from these complex and multifaceted data. In addition, we show how the model can be used to analyze sequences that contribute to assessment of mortality likelihood.

show abstract

Nonstationary multivariate Gaussian processes for electronic health records

Cited by 12 publications

References 34 publications

Bayesian Inference in High-Dimensional Time-Serieswith the Orthogonal Stochastic Linear Mixing Model

Bayesian Inference in High-Dimensional Time-Serieswith the Orthogonal Stochastic Linear Mixing Model

Dynamic modeling of hospitalized COVID-19 patients reveals disease state–dependent risk factors

Unsupervised Probabilistic Models for Sequential Electronic Health Records

Contact Info

Product

Resources

About