Case Studies in Applied Bayesian Data Science: Cirm Jean-Morlet Chair, Fall 2018

ISBN-13: 9783030425524 / Angielski / Miękka / 2020 / 420 str.

Kerrie Mengersen; Pierre Pudlo; Christian P. Robert

Spis treści

- Part I Surveys. - Introduction. - A Survey of Bayesian Statistical Approaches for Big Data. - Bayesian Neural Networks: An Introduction and Survey. - Markov Chain Monte Carlo Algorithms for Bayesian Computation, a Survey and Some Generalisation. - Bayesian Variable Selection. - Bayesian Computation with Intractable Likelihoods. - Part II Real World Case Studies in Health. - A Bayesian Hierarchical Approach to Jointly Model Cortical Thickness and Covariance Networks. - Bayesian Spike Sorting: Parametric and Nonparametric Multivariate Gaussian MixtureModels. - Spatio-Temporal Analysis of Dengue Fever in Makassar Indonesia: A Comparison of Models Based on CARBayes. - A Comparison of Bayesian Spatial Models for Cancer Incidence at a Small Area Level: Theory and Performance. - An Ensemble Approach to Modelling the Combined Effect of Risk Factors on Age at Parkinson’s Disease Onset. - Workplace Health and Workplace Wellness: Synergistic or Disconnected?. - Bayesian Modelling to Assist Inference on Health Outcomes in Occupational Health Surveillance. Part III Real World Case Studies in Ecology. - Bayesian Networks for Understanding Human-Wildlife Conflict in Conservation. - Bayesian Learning of Biodiversity Models Using Repeated Observations. - Thresholds of Coral Cover That Support Coral Reef Biodiversity. - Application of Bayesian Mixture Models to Satellite Images and Estimating the Risk of Fire-Ant Incursion in the Identified Geographical Cluster.

Blurb

Presenting a range of substantive applied problems within Bayesian Statistics along with their Bayesian solutions, this book arises from a research program at CIRM in France in the second semester of 2018, which supported Kerrie Mengersen as a visiting Jean-Morlet Chair and Pierre Pudlo as the local Research Professor.

The field of Bayesian statistics has exploded over the past thirty years and is now an established field of research in mathematical statistics and computer science, a key component of data science, and an underpinning methodology in many domains of science, business and social science. Moreover, while remaining naturally entwined, the three arms of Bayesian statistics, namely modelling, computation and inference, have grown into independent research fields.While the research arms of Bayesian statistics continue to grow in many directions, they are harnessed when attention turns to solving substantive applied problems. Each such problem set has its own challenges and hence draws from the suite of research a bespoke solution.

The book will be useful for both theoretical and applied statisticians, as well as practitioners, to inspect these solutions in the context of the problems, in order to draw further understanding, awareness and inspiration.