Billard / Diday Symbolic Data Analysis

Conceptual Statistics and Data Mining
Erscheinungsjahr 2012
ISBN: 978-0-470-09017-6
Verlag: John Wiley & Sons
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 330 Seiten, E-Book

Reihe: Wiley Series in Computational Statistics

With the advent of computers, very large datasets have becomeroutine. Standard statistical methods don't have the power orflexibility to analyse these efficiently, and extract the requiredknowledge. An alternative approach is to summarize a large datasetin such a way that the resulting summary dataset is of a manageablesize and yet retains as much of the knowledge in the originaldataset as possible. One consequence of this is that the data mayno longer be formatted as single values, but be represented bylists, intervals, distributions, etc. The summarized data havetheir own internal structure, which must be taken into account inany analysis.
This text presents a unified account of symbolic data, how theyarise, and how they are structured. The reader is introduced tosymbolic analytic methods described in the consistent statisticalframework required to carry out such a summary and subsequentanalysis.
* Presents a detailed overview of the methods and applications ofsymbolic data analysis.
* Includes numerous real examples, taken from a variety ofapplication areas, ranging from health and social sciences, toeconomics and computing.
* Features exercises at the end of each chapter, enabling thereader to develop their understanding of the theory.
* Provides a supplementary website featuring links to downloadthe SODAS software developed exclusively for symbolic dataanalysis, data sets, and further material.
Primarily aimed at statisticians and data analysts, SymbolicData Analysis is also ideal for scientists working on problemsinvolving large volumes of data from a range of disciplines,including computer science, health and the social sciences. Thereis also much of use to graduate students of statistical dataanalysis courses.

Autoren/Hrsg.

Billard, Lynne

Diday, Edwin

Inhaltsverzeichnis

1. Introduction.
References.
2. Symbolic Data.
2.1 Symbolic and Classical Data.
2.2 Categories, Concepts and Symbolic Objects.
2.3 Comparison of Symbolic and Classical Analysis.
3. Basic Descriptive Statistics: One Variate.
3.1 Some Preliminaries.
3.2 Multi-valued Variables.
3.3 Interval-valued Variables.
3.4 Multi-valued Modal variables.
3.5 Interval-valued Modal Variables.
4. Descriptive Statistics: Two or More Variates.
4.1 Multi-valued Variables.
4.2 Interval-valued Variables.
4.3 Modal Multi-valued Variables.
4.4 Modal Interval-valued Variables.
4.5 Baseball Interval-valued Dataset.
4.6 Measures of Dependence.
5. Principal Component Analysis.
5.1 Vertices Method.
5.2 Centers Method.
5.3 Comparison of the Methods.
6. Regression Analysis.
6.1 Classical Multiple Regression Model.
6.2 Multi-valued Variables.
6.3 Interval-valued Variables.
6.4 Histogram-valued Variables.
6.5 Taxonomy Variables.
6.6 Hierarchical Variables.
7. Cluster Analysis.
7.1 Dissimilarity and Distance Measures.
7.2 Clustering Structures.
7.3 Partitions.
7.4 Hierarchy-Divisive Clustering.
7.5 Hierarchy-Pyramid Clusters.
Data Index.
Author Index.
Subject Index.

Über Autor(innen)

LynneBillard is a multi award winning University Professor ofStatistics at the University of Georgia, USA. Her areas of interestinclude epidemic theory, AIDS, time series, sequential analysis,and symbolic data. A former President of the American StatisticalAssociation as well as the ENAR Regional President andInternational President of the International Biometric Society,Professor Billard has co-edited 6 books, published over150 papersand been actively involved in many statistical societies andnational committees.
Edwin Diday is a Professor in Computer Science andMathematics, at the Université Paris Dauphine, France. He isthe author or editor of 14 previous books. He is also the founderof the symbolic data analysis field, and has led numerousinternational research teams in the area.