Muutke küpsiste eelistusi

Statistical Modeling With R: a dual frequentist and Bayesian approach for life scientists [Pehme köide]

4.00/5 (2 hinnangut Goodreads-ist)
(Professor of Ecology, Universidad de la República, Centro Universitario Regional del Este, Uruguay)
  • Formaat: Paperback / softback, 480 pages, kõrgus x laius x paksus: 246x189x22 mm, kaal: 1044 g
  • Ilmumisaeg: 02-Nov-2022
  • Kirjastus: Oxford University Press
  • ISBN-10: 0192859021
  • ISBN-13: 9780192859020
Teised raamatud teemal:
Statistical Modeling With R: a dual frequentist and Bayesian approach for life scientistsSuurem pilt
  • Formaat: Paperback / softback, 480 pages, kõrgus x laius x paksus: 246x189x22 mm, kaal: 1044 g
  • Ilmumisaeg: 02-Nov-2022
  • Kirjastus: Oxford University Press
  • ISBN-10: 0192859021
  • ISBN-13: 9780192859020
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9780192859020.html
Märksõnad:
To date, statistics has tended to be neatly divided into two theoretical approaches or frameworks: frequentist (or classical) and Bayesian. Scientists typically choose the statistical framework to analyse their data depending on the nature and complexity of the problem, and based on their personal views and prior training on probability and uncertainty. Although textbooks and courses should reflect and anticipate this dual reality, they rarely do so. This accessible textbook explains, discusses, and applies both the frequentist and Bayesian theoretical frameworks to fit the different types of statistical models that allow an analysis of the types of data most commonly gathered by life scientists. It presents the material in an informal, approachable, and progressive manner suitable for readers with only a basic knowledge of calculus and statistics.

Statistical Modeling with R is aimed at senior undergraduate and graduate students, professional researchers, and practitioners throughout the life sciences, seeking to strengthen their understanding of quantitative methods and to apply them successfully to real world scenarios, whether in the fields of ecology, evolution, environmental studies, or computational biology.

Arvustused

A book that should attract curious minds of various backgrounds, knowledge, and expertise in statistics, as well as work nicely to support an enthusiastic teacher of statistical modeling. I thus recommend this book most enthusiastically. * Christian P. Robert, Journal of the American Statistical Association * The book is a novel contribution to the literature on statistical modelling, it has my highest endorsement, and I look forward to using it in future graduate courses on applied statistics. * Lars Rönnegård, Dalarna University *

Preface vii
PART I The Conceptual Basis for Fitting Statistical Models
1 General Introduction
3(6)
1.1 The purpose of statistics
3(1)
1.2 Statistics in a schizophrenic state?
4(1)
1.3 How is this book organized?
4(2)
1.4 How to use this book
6(3)
References
7(2)
2 Statistical Modeling: A short historical background
9(10)
2.1 What is a statistical model?
9(1)
2.2 What is this thing called probability?
10(3)
2.3 Linking probability with statistics
13(2)
2.4 The early Bayesian demise during the 1930s
15(4)
References
17(2)
3 Estimating Parameters: The main purpose of statistical inference
19(40)
3.1 Introduction
19(1)
3.2 Least squares: A theory of errors and the normal distribution
20(1)
3.3 Maximum likelihood
20(9)
3.3.1 The basic concepts
20(2)
3.3.2 Obtaining maximum likelihood estimates
22(5)
3.3.3 Using maximum likelihood estimates in statistical inference
27(2)
3.4 Bayesian parameter estimation: The basics
29(7)
3.5 Bayesian methods: Markov chain Monte Carlo to the rescue
36(9)
3.6 Quality control for the algorithms of Bayesian methods
45(1)
3.7 More general MCMC variations: Metropolis-Hastings and Gibbs algorithms
46(2)
3.8 Recent advances in Bayesian methods: Hamiltonian Monte Carlo
48(2)
3.9 Bayesian hypothesis tests
50(1)
3.10 Summary of the main differences between maximum likelihood and Bayesian methods
51(8)
References
54(5)
PART II Applying the Generalized Linear Model to Varied Data Types
4 The General Linear Model I: Numerical explanatory variables
59(42)
4.1 Introduction
59(1)
4.2 The lognormal distribution and its relation to the general linear model
60(1)
4.3 Simple linear regression: One continuous explanatory variable
61(3)
4.4 Simple linear regression: Frequentist fitting
64(2)
4.5 Tools for model validation in frequentist statistics
66(3)
4.6 Simple linear regression: Bayesian fitting
69(9)
4.7 Tools for model validation in Bayesian statistics
78(2)
4.8 Multiple linear regression: More than one numerical explanatory variable
80(4)
4.9 Multiple linear regression: Frequentist fitting
84(2)
4.10 The importance of standardizing explanatory variables
86(4)
4.11 Polynomial regression
90(1)
4.12 Multiple linear regression: Bayesian fitting
91(7)
4.13 Problems
98(3)
References
98(3)
5 The General Linear Model II: Categorical explanatory variables
101(38)
5.1 Introduction
101(1)
5.2 Student's t test: One categorical explanatory variable with two groups
101(5)
5.3 The t-test: Frequentist fitting
106(4)
5.4 The t-test: Bayesian fitting
110(5)
5.5 Viewing one-way analysis of variance as a multiple regression
115(5)
5.6 One-way analysis of variance: Frequentist fitting
120(4)
5.7 One-way analysis of variance: Bayesian fitting
124(6)
5.8 A posteriori tests in frequentist models
130(4)
5.9 A posteriori tests in Bayesian models?
134(2)
5.10 Problems
136(3)
References
137(2)
6 The General Linear Model III: Interactions between explanatory variables
139(30)
6.1 Introduction
139(1)
6.2 Factorial analysis of variance
139(6)
6.3 Factorial analysis of variance: Frequentist fitting
145(4)
6.4 Factorial analysis of variance: Bayesian fitting
149(7)
6.5 Analysis of covariance: Mixing continuous and categorical explanatory variables
156(2)
6.6 Analysis of covariance: Frequentist fitting
158(4)
6.7 Analysis of covariance: Bayesian fitting
162(6)
6.8 Problems
168(1)
References
168(1)
7 Model Selection: One, two, and more models fitted to the data
169(20)
7.1 Introduction
169(1)
7.2 The problem of model selection: Parsimony in statistics
170(2)
7.3 Model selection criteria in the frequentist framework: AIC
172(4)
7.4 Model selection criteria in the Bayesian framework: DIC and WAIC
176(2)
7.5 The posterior predictive distribution and posterior predictive checks
178(4)
7.6 Now back to the WAIC and LOO-CV
182(3)
7.7 Prior predictive distributions: A relatively "new" kid on the block
185(4)
References
186(3)
8 The Generalized Linear Model
189(12)
8.1 Introduction
189(1)
8.2 What are GLMs made of?
189(4)
8.3 Fitting GLMs
193(1)
8.4 Goodness of fit in GLMs
194(4)
8.5 Statistical significance of GLM
198(3)
References
198(3)
9 When the Response Variable is Binary
201(34)
9.1 Introduction
201(1)
9.2 Key concepts for binary GLMs: Odds, log odds, and additional link functions
202(1)
9.3 Fitting binary GLMs
203(4)
9.4 Ungrouped binary GLM: Frequentist fitting
207(6)
9.5 Further issues about validating binary GLMs
213(3)
9.6 Ungrouped binary GLMs: Bayesian fitting
216(11)
9.7 Grouped binary GLMs
227(6)
9.8 Problems
233(2)
References
233(2)
10 When the Response Variable is a Count, Often with Many Zeros
235(36)
10.1 Introduction
235(2)
10.2 Over-dispersion: A common problem with many causes and some solutions
237(2)
10.3 Plant species richness and geographical variables
239(15)
10.3.1 Frequentist fitting of the count GLM
242(5)
10.3.2 Bayesian fitting of count GLMs
247(7)
10.4 Modeling of counts with an excess of zeros: Zero-inflated and hurdle models
254(14)
10.4.1 Frequentist fitting of a zero-inflated model
256(5)
10.4.2 Bayesian fitting of a zero-augmented model
261(7)
10.5 Problems
268(3)
References
269(2)
11 Further Issues Involved in the Modeling of Counts
271(22)
11.1 "The more you search, the more you find"
271(1)
11.2 Log-linear models as count GLMs
272(3)
11.3 Frequentist fitting of a log-linear model
275(9)
11.4 Bayesian fitting of a log-linear model
284(7)
11.5 Problems
291(2)
References
292(1)
12 Models for Positive, Real-Valued Response Variables: Proportions and others
293(34)
12.1 Introduction
293(1)
12.2 Modeling proportions
293(2)
12.3 Plant cover, grazing, and productivity
295(2)
12.4 Frequentist fitting of a GLM on proportions
297(6)
12.5 Bayesian fitting of a GLM on proportions
303(8)
12.6 Modeling positive, real-valued response variables
311(1)
12.7 Predicting tree seedling biomass
312(2)
12.8 Frequentist fitting of a gamma GLM
314(3)
12.9 Bayesian fitting of a gamma GLM
317(2)
12.10 Other related yet important cases of positive, real-valued response variables
319(1)
12.11 Problems
320(7)
References
321(2)
Approaches to Defining Priors
323(4)
PART III Incorporating Experimental and Survey Design Using Mixed Models
13 Accounting for Structure in Mixed/Hierarchical Models
327(46)
13.1 Introduction
327(2)
13.2 Fixed effects and random effects in the frequentist framework
329(3)
13.3 Defining mixed effects models
332(3)
13.4 Problems and inconsistencies with the definition of random effects
335(1)
13.5 Population-level and group-level effects in Bayesian hierarchical models
336(3)
13.6 Fitting mixed models in the frequentist framework
339(14)
13.7 Statistical significance and model selection in frequentist mixed models
353(3)
13.8 The shrinkage or borrowing strength effect in mixed models
356(2)
13.9 Fitting mixed models in the Bayesian framework
358(12)
13.10 Problems
370(3)
References
371(2)
14 Experimental Design in the Life Sciences: The basics
373(20)
14.1 Introduction
373(1)
14.2 The basic principles of experimental design
374(2)
14.3 Surveys and observational studies
376(1)
14.4 The main types of experimental design used in the life sciences
376(11)
14.4.1 Factorial design
377(1)
14.4.2 Randomized block design
378(2)
14.4.3 Split-plot design
380(2)
14.4.4 Nested design
382(2)
14.4.5 Repeated measures design
384(1)
14.4.6 Crossover design
385(2)
14.5 How many samples should we take?
387(6)
References
390(3)
15 Mixed Hierarchical Models and Experimental Design Data
393(60)
15.1 Introduction
393(1)
15.2 Binary GLMM with a randomized block design
394(22)
15.2.1 Binary GLMM with a randomized block design: Frequentist models
398(9)
15.2.2 Binary GLMM with a randomized block design: Bayesian models
407(9)
15.3 Gaussian GLMM with a repeated measures design
416(12)
15.3.1 Gaussian GLMM with a repeated measures design: Frequentist models
420(3)
15.3.2 Gaussian GLMM with a repeated measures design: Bayesian models
423(5)
15.4 Beta GLMM with a split-plot design
428(21)
15.4.1 Beta GLMM with a split-plot design: Frequentist model
432(7)
15.4.2 Beta GLMM with a split-plot design: Bayesian model
439(10)
15.5 Problems
449(4)
References
449(2)
Afterword
451(2)
Appendix A List of R Packages Used in This Book 453(2)
Appendix B Exploring and Describing the Evidence in Graphics (only available online at www.oup.com/companion/InchaustiSMWR)
Appendix C Using R and RStudio: The Bare-Bones Basics (only available online at www.oup.com/companion/InchaustiSMWR)
Index 455
Pablo Inchausti is Professor of Ecology at the Universidad de la República, Centro Universitario Regional del Este, Uruguay. He is the co-editor of the influential and highly-cited book Biodiversity and Ecosystem Functioning: synthesis and perspectives (OUP, 2002) and has been successfully teaching statistics and mathematical modelling to students of the life and social sciences for over 15 years.