Some Statistical Problems Related to Big-Data

来源 :上海交通大学 | 被引量 : 0次 | 上传用户：JK0803_liuchao

【摘要】

：

【作者】

：

Boaz Nadler Jonathan Rosenblatt

【机构】

：

Weizmann Institute of Science,Rehovot,76100,Israel

【出处】

：

上海交通大学

【发表日期】

：

2016年4期

【关键词】

：

Distributed learning M-estimation Empirical risk minimization Big data High orde

下载到本地 , 更方便阅读

下载此文赞助VIP

声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架

论文部分内容阅读

　　One key challenge with big-data is that it may not fit or cannot be processed in a single machine.A common approach to statistical learning in this setting is to randomly split the data among m machines,let each compute its own estimator which is then sent to a central node for merging.

其他文献

A New Weighted Partial Likelihood Method for Estimating Marginal Structural Hazard Models

Parameters in marginal structural Cox models can be estimated through an inverse probability of treatment and censoring(IPTC)weighted Cox partial likelihood.

会议

Case-base samplingEvent history analysisInverse probability of treatment and c

Nonparametric Dynamic Screening System for Monitoring Correlated Longitudinal Data

In many applications,including disease early detection and prevention,and performance evaluation of airplanes and other durable products,we need to sequentially monitor the longitudinal pattern of cer

会议

Cholesky decompositionDynamic screeningLongitudinal dataStatistical process c

Removing Unwanted Variation with Replicates and Negative Controls

High-throughput biological data,such as microarray data and gene sequencing data,are plagued by unwanted variation – systematic errors introduced by variations in experimental conditions such as tempe

会议

genomicslatent factorsnegative controlsunwanted variation

Leveraging Bayesian Network meta-analysis to improve drug development

The extent of phase 3 failures contributes significantly to the increasing costs of drug development.

会议

drug developmentBayesian network meta-analysis

Functional data analysis in application to dose-titration studies

Functional data analysis is powerful in analyzing data that can be represented by curves or images.

会议

functional datatesting of functional datafunctional mixed effect modelclinica

Incorporating historical data in Bayesian phase Ⅰ trial design:Evaluating the similarity in dose-tox

In oncology area,following a phase Ⅰ dose-finding trial completed in a certain population of patients,further phase Ⅰ trials are often conducted to determine the maximum tolerated dose(MTD)for differe

会议

Phase Ⅰ clinical trialdose-findingBayesian trial designhistorical databorrow

Acceleration of convergence of the alternating least squares algorithm for mixed measurement level m

The alternating least squares(ALS)algorithm is a most popular iterative computational method for complex data,such as mixed measurement levels data.

会议

The ALS algorithmThe vector ε algorithmAcceleration of convergenceRe-starting

Analysis of Error-Contaminated Survival Data

Covariate measurement error has attracted extensive interest in survival analysis.Since Prentice(1982),a large number of inference methods have been developed to handle error-contaminated data,and mos

会议

Additive hazards modelMeasurement errorSurvival analysis

Partially Linear Single-Index Regression for Accelerated Failure Time Models

In survival analysis,Accelerated Failure Time(AFT)models are very useful in modeling the relationship between failure times and associated covariates,where covariate effects are usually assumed to app

会议

Accelerated Failure Time ModelMisspecificationPartially Linear Single Index Mo

A joint selection of designs and models for optimal forecasting in possibly misspecified polynomial

The classical optimal design methods have the capability of determining designs to achieve estimation or prediction efficiency in situations where the working model is correctly specified.

会议

Information criterionmean-squared prediction errormodel selectionoptimality c

Some Statistical Problems Related to Big-Data

其他学术论文