Curved exponential family models for social networks

The aim of the research presented in the paper is to explore the inter-relations between care processes and personal social networks as social capital in the light of the changing family models. Research of interdependence of care, social capital and family models is based on the idea of family practices suggested by Morgan. The main research question is what family practices of various family models create such social capital that ensure caring for its' members? The research hypothesis is that participation in various activities together with family members and persons beyond nuclear and extended family create dense social networks of caregivers. The analysis is based on data of representative quantitative survey carried out in Lithuania between 2011 November-2012 May within the ESF supported research project "Trajectories of family models and social networks: intergenerational perspective". Research results only partly support this hypothesis: particular family practices create networks of caregivers, but in order to involve particular persons into network of caregivers, different family practices in various family models are needed. Usually, inhabitants of Lithuania primarily expect to receive care from persons who depend to nuclear family created through marriage and extended family arisen from this relation. But persons from whom it is expected to receive care and care received differ in Lithuania. In reality, caregivers usually are children in families with children and parents in families without children. Family practices that create social networks of caregivers, and are common for all family models include annual feasts like Christmas Eve, Christmas, Easter, All Soul's Day, New Year party, Mother's Day. Various family practices differently impacting creation social networks of caregivers for different family models but usually its include joint dinner daily, Sunday lunch together, vacations with family, communication face-toface, by the telephone or Internet, consultations on important decision-making, All Soul's Day feast, Christmas celebration, Mother's Day, Gatherings of relatives, Birthday, Name-day feast, visiting cultural event together.

Download Full-text

Nonstandard conditionally specified models for nonignorable missing data

Proceedings of the National Academy of Sciences ◽

10.1073/pnas.1815563117 ◽

2020 ◽

Vol 117 (32) ◽

pp. 19045-19053

Author(s):

Alexander M. Franks ◽

Edoardo M. Airoldi ◽

Donald B. Rubin

Keyword(s):

Systems Biology ◽

Missing Data ◽

Joint Distribution ◽

Exponential Family ◽

Nonignorable Missing Data ◽

Data Analyses ◽

Missingness Mechanism ◽

Conditionally Specified Models ◽

Nonignorable Missing ◽

Family Models

Data analyses typically rely upon assumptions about the missingness mechanisms that lead to observed versus missing data, assumptions that are typically unassessable. We explore an approach where the joint distribution of observed data and missing data are specified in a nonstandard way. In this formulation, which traces back to a representation of the joint distribution of the data and missingness mechanism, apparently first proposed by J. W. Tukey, the modeling assumptions about the distributions are either assessable or are designed to allow relatively easy incorporation of substantive knowledge about the problem at hand, thereby offering a possibly realistic portrayal of the data, both observed and missing. We develop Tukey’s representation for exponential-family models, propose a computationally tractable approach to inference in this class of models, and offer some general theoretical comments. We then illustrate the utility of this approach with an example in systems biology.

Download Full-text

A Simple and Accurate Method for Approximate Conditional Inference Applied to Exponential Family Models

Journal of the Royal Statistical Society Series B (Methodological) ◽

10.1111/j.2517-6161.1996.tb02074.x ◽

1996 ◽

Vol 58 (1) ◽

pp. 177-188

Author(s):

Richard P. Waterman ◽

Bruce G. Lindsay

Keyword(s):

Exponential Family ◽

Accurate Method ◽

Conditional Inference ◽

Family Models

Download Full-text

Proper and non-informative conjugate priors for exponential family models

Bayesian Theory and Applications ◽

10.1093/acprof:oso/9780199695607.003.0019 ◽

2013 ◽

pp. 395-408

Author(s):

E GUTIéRREZ - PEñA ◽

M MENDOZA

Keyword(s):

Exponential Family ◽

Conjugate Priors ◽

Family Models

Download Full-text

Variational Inference over Nonstationary Data Streams for Exponential Family Models

Mathematics ◽

10.3390/math8111942 ◽

2020 ◽

Vol 8 (11) ◽

pp. 1942

Author(s):

Andrés R. Masegosa ◽

Darío Ramos-López ◽

Antonio Salmerón ◽

Helge Langseth ◽

Thomas D. Nielsen

Keyword(s):

Bayesian Inference ◽

Variational Methods ◽

Latent Variables ◽

Exponential Family ◽

Concept Drift ◽

Variational Inference ◽

Model Parameters ◽

Time Step ◽

Data Set ◽

Family Models

In many modern data analysis problems, the available data is not static but, instead, comes in a streaming fashion. Performing Bayesian inference on a data stream is challenging for several reasons. First, it requires continuous model updating and the ability to handle a posterior distribution conditioned on an unbounded data set. Secondly, the underlying data distribution may drift from one time step to another, and the classic i.i.d. (independent and identically distributed), or data exchangeability assumption does not hold anymore. In this paper, we present an approximate Bayesian inference approach using variational methods that addresses these issues for conjugate exponential family models with latent variables. Our proposal makes use of a novel scheme based on hierarchical priors to explicitly model temporal changes of the model parameters. We show how this approach induces an exponential forgetting mechanism with adaptive forgetting rates. The method is able to capture the smoothness of the concept drift, ranging from no drift to abrupt drift. The proposed variational inference scheme maintains the computational efficiency of variational methods over conjugate models, which is critical in streaming settings. The approach is validated on four different domains (energy, finance, geolocation, and text) using four real-world data sets.

Download Full-text