|
Showing 1 - 7 of
7 matches in All Departments
This volume conveys some of the surprises, puzzles and success
stories in high-dimensional and complex data analysis and related
fields. Its peer-reviewed contributions showcase recent advances in
variable selection, estimation and prediction strategies for a host
of useful models, as well as essential new developments in the
field. The continued and rapid advancement of modern technology now
allows scientists to collect data of increasingly unprecedented
size and complexity. Examples include epigenomic data, genomic
data, proteomic data, high-resolution image data, high-frequency
financial data, functional and longitudinal data, and network data.
Simultaneous variable selection and estimation is one of the key
statistical problems involved in analyzing such big and complex
data. The purpose of this book is to stimulate research and foster
interaction between researchers in the area of high-dimensional
data analysis. More concretely, its goals are to: 1) highlight and
expand the breadth of existing methods in big data and
high-dimensional data analysis and their potential for the
advancement of both the mathematical and statistical sciences; 2)
identify important directions for future research in the theory of
regularization methods, in algorithmic development, and in
methodologies for different application areas; and 3) facilitate
collaboration between theoretical and subject-specific researchers.
This volume conveys some of the surprises, puzzles and success
stories in high-dimensional and complex data analysis and related
fields. Its peer-reviewed contributions showcase recent advances in
variable selection, estimation and prediction strategies for a host
of useful models, as well as essential new developments in the
field. The continued and rapid advancement of modern technology now
allows scientists to collect data of increasingly unprecedented
size and complexity. Examples include epigenomic data, genomic
data, proteomic data, high-resolution image data, high-frequency
financial data, functional and longitudinal data, and network data.
Simultaneous variable selection and estimation is one of the key
statistical problems involved in analyzing such big and complex
data. The purpose of this book is to stimulate research and foster
interaction between researchers in the area of high-dimensional
data analysis. More concretely, its goals are to: 1) highlight and
expand the breadth of existing methods in big data and
high-dimensional data analysis and their potential for the
advancement of both the mathematical and statistical sciences; 2)
identify important directions for future research in the theory of
regularization methods, in algorithmic development, and in
methodologies for different application areas; and 3) facilitate
collaboration between theoretical and subject-specific researchers.
The objective of this book is to compare the statistical
properties of penalty and non-penalty estimation strategies for
some popular models. Specifically, it considers the full model,
submodel, penalty, pretest and shrinkage estimation techniques for
three regression models before presenting the asymptotic properties
of the non-penalty estimators and their asymptotic distributional
efficiency comparisons. Further, the risk properties of the
non-penalty estimators and penalty estimators are explored through
a Monte Carlo simulation study. Showcasing examples based on real
datasets, the book will be useful for students and applied
researchers in a host of applied fields.
The book's level of presentation and style make it accessible to
a broad audience. It offers clear, succinct expositions of each
estimation strategy. More importantly, it clearly describes how to
use each estimation strategy for the problem at hand. The book is
largely self-contained, as are the individual chapters, so that
anyone interested in a particular topic or area of application may
read only that specific chapter. The book is specially designed for
graduate students who want to understand the foundations and
concepts underlying penalty and non-penalty estimation and its
applications. It is well-suited as a textbook for senior
undergraduate and graduate courses surveying penalty and
non-penalty estimation strategies, and can also be used as a
reference book for a host of related subjects, including courses on
meta-analysis. Professional statisticians will find this book to be
a valuable reference work, since nearly all chapters are
self-contained.
This volume features selected, refereed papers on various aspects
of statistics, matrix theory and its applications to statistics, as
well as related numerical linear algebra topics and numerical
solution methods, which are relevant for problems arising in
statistics and in big data. The contributions were originally
presented at the 25th International Workshop on Matrices and
Statistics (IWMS 2016), held in Funchal (Madeira), Portugal on June
6-9, 2016. The IWMS workshop series brings together statisticians,
computer scientists, data scientists and mathematicians, helping
them better understand each other's tools, and fostering new
collaborations at the interface of matrix theory and statistics.
This volume features selected, refereed papers on various aspects
of statistics, matrix theory and its applications to statistics, as
well as related numerical linear algebra topics and numerical
solution methods, which are relevant for problems arising in
statistics and in big data. The contributions were originally
presented at the 25th International Workshop on Matrices and
Statistics (IWMS 2016), held in Funchal (Madeira), Portugal on June
6-9, 2016. The IWMS workshop series brings together statisticians,
computer scientists, data scientists and mathematicians, helping
them better understand each other's tools, and fostering new
collaborations at the interface of matrix theory and statistics.
This volume contains the proceedings of the International Workshop
on Perspectives on High-dimensional Data Analysis II, held May
30-June 1, 2012, at the Centre de Recherches Mathematiques,
Universite de Montreal, Montreal, Quebec, Canada. This book
collates applications and methodological developments in
high-dimensional statistics dealing with interesting and
challenging problems concerning the analysis of complex,
high-dimensional data with a focus on model selection and data
reduction. The chapters contained in this book deal with submodel
selection and parameter estimation for an array of interesting
models. The book also presents some surprising results on
high-dimensional data analysis, especially when signals cannot be
effectively separated from the noise, it provides a critical
assessment of penalty estimation when the model may not be sparse,
and it suggests alternative estimation strategies. Readers can
apply the suggested methodologies to a host of applications and
also can extend these methodologies in a variety of directions.
This volume conveys some of the surprises, puzzles and success
stories in big data analysis and related fields. This book is
co-published with the Centre de Recherches Mathematiques.
|
You may like...
Loot
Nadine Gordimer
Paperback
(2)
R398
R330
Discovery Miles 3 300
Poor Things
Emma Stone, Mark Ruffalo, …
DVD
R449
R329
Discovery Miles 3 290
|