A Distributed and Integrated Method of Moments for High-Dimensional Correlated Data Analysis

Hector, Emily C.; Song, Peter X.‐K.

doi:10.1080/01621459.2020.1736082

Cited by 17 publications

(6 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar questions have been studied in the literature (Hector & Song, 2021Kleiner et al, 2014;Mackey et al, 2015;Shamir et al, 2014;. Many of them focus on parallel computing involving iterative communications until convergence (Maclaurin & Adams, 2015;Scaman et al, 2018).…”

Section: Introductionmentioning

confidence: 80%

CEDAR: Communication Efficient Distributed Analysis for Regressions

Chang

Long

2022

Biometrics

View full text Add to dashboard Cite

Electronic health records (EHRs) offer great promises for advancing precision medicine and, at the same time, present significant analytical challenges. Particularly, it is often the case that patient‐level data in EHRs cannot be shared across institutions (data sources) due to government regulations and/or institutional policies. As a result, there are growing interests about distributed learning over multiple EHRs databases without sharing patient‐level data. To tackle such challenges, we propose a novel communication efficient method that aggregates the optimal estimates of external sites, by turning the problem into a missing data problem. In addition, we propose incorporating posterior samples of remote sites, which can provide partial information on the missing quantities and improve efficiency of parameter estimates while having the differential privacy property and thus reducing the risk of information leaking. The proposed approach, without sharing the raw patient level data, allows for proper statistical inference. We provide theoretical investigation for the asymptotic properties of the proposed method for statistical inference as well as differential privacy, and evaluate its performance in simulations and real data analyses in comparison with several recently developed methods.

show abstract

Section: Introductionmentioning

confidence: 80%

CEDAR: Communication Efficient Distributed Analysis for Regressions

Chang

Long

2022

Biometrics

View full text Add to dashboard Cite

show abstract

“…where (W ) k,k denotes the rows and columns of W corresponding to subsets D k and D k respectively, for any positive semi-definite weight matrix W . This approach has been successfully employed by others (Bai et al, 2012;Hector and Song, 2021) although never with a MSP or censored (composite) likelihood, and has connections to weighted composite likelihood (Le Cessie and van Houwelingen, 1994;Nott and Rydén, 1999;Kuk, 2007;Joe and Lee, 2009;Zhao and Joe, 2009;Sang and Genton, 2014;Castruccio et al, 2016). Under mild regularity conditions (Newey and McFadden, 1994), θ GM M is a consistent estimator of θ and asymptotically normally distributed as n → ∞:…”

Section: Define the Stacking Operation {Amentioning

confidence: 99%

“…The iterative minimization in (5) remains computationally burdensome for large d because the censored composite score function of d k pairs must be evaluated at each iteration of the minimization. Fortunately, this iterative procedure may be altogether bypassed through the closed-form meta-estimator derived by Hector and Song (2021):…”

Section: Implementation: a Meta-estimatormentioning

confidence: 99%

Distributed Inference for Spatial Extremes Modeling in High Dimensions

Hector¹,

Reich²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Extreme environmental events frequently exhibit spatial and temporal dependence. These data are often modeled using max stable processes (MSPs). MSPs are computationally prohibitive to fit for as few as a dozen observations, with supposed computationally-efficient approaches like the composite likelihood remaining computationally burdensome with a few hundred observations. In this paper, we propose a spatial partitioning approach based on local modeling of subsets of the spatial domain that delivers computationally and statistically efficient inference. Marginal and dependence parameters of the MSP are estimated locally on subsets of observations using censored pairwise composite likelihood, and combined using a modified generalized method of moments procedure. The proposed distributed approach is extended to estimate spatially varying coefficient models to deliver computationally efficient modeling of spatial variation in marginal parameters. We demonstrate consistency and asymptotic normality of estimators, and show empirically that our approach leads to a surprising reduction in bias of parameter estimates over a full data approach. We illustrate the flexibility and practicability of our approach through simulations and the analysis of streamflow data from the U.S. Geological Survey.

show abstract

“…Under mild regularity conditions, √ n k U k (β 0 ) is asymptotically normally distributed with mean 0 and variance φ −1 0 j(β 0 ), which is consistently estimated by φ −1 k J k . Letting Φ p be the cumulative distribution function of the p-variate standard Normal distribution, the asymptotic confidence estimating function (CEF), as defined in Hector and Song (2020a)…”

Section: Combine Algorithmmentioning

confidence: 99%

Parallel-and-stream accelerator for computationally fast supervised learning

Hector¹,

Li²,

Song³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Two dominant distributed computing strategies have emerged to overcome the computational bottleneck of supervised learning with big data: parallel data processing in the MapReduce paradigm and serial data processing in the online streaming paradigm. Despite the two strategies' common divide-and-combine approach, they differ in how they aggregate information, leading to different trade-offs between statistical and computational performance. In this paper, we propose a new hybrid paradigm, termed a Parallel-and-Stream Accelerator (PASA), that uses the strengths of both strategies for computationally fast and statistically efficient supervised learning. PASA's architecture nests online streaming processing into each distributed and parallelized data process in a MapReduce framework. PASA leverages the advantages and mitigates the disadvantages of both the MapReduce and online streaming approaches to deliver a more flexible paradigm satisfying practical computing needs. We study the analytic properties and computational complexity of PASA, and detail its implementation for two key statistical learning tasks. We illustrate its performance through simulations and a large-scale data example building a prediction model for online purchases from advertising data.

show abstract

A Distributed and Integrated Method of Moments for High-Dimensional Correlated Data Analysis

Cited by 17 publications

References 46 publications

CEDAR: Communication Efficient Distributed Analysis for Regressions

CEDAR: Communication Efficient Distributed Analysis for Regressions

Distributed Inference for Spatial Extremes Modeling in High Dimensions

Parallel-and-stream accelerator for computationally fast supervised learning

Contact Info

Product

Resources

About