GEOS36502/EVOL 33002: Paleobiological Modeling and Analysis 2 (Multivariate Analysis) [Foote]

Winter 2013

Syllabus

List of readings (see Syllabus for dates)

General Bibliography for Multivariate Analysis

LECTURE OUTLINES (to be posted as the quarter progresses)

1. Introduction and overview; data transformations

2. Bivariate analysis

3. Measuring Similarity and Distance

4. Cluster Analysis

5. Introduction to Ordination

6. Linear Algebra I

7. Linear Algebra II

8. Linear Algebra III; Principal Component Analysis I

9. Principal Component Analysis II

Selected Figures on Principal Component Analysis

10. Miscellaneous Eigenvector Methods

11. R-mode Factor Analysis

R-mode Factor Analysis: Selected Figures

12. Q-mode Eigenvector Analysis; SVD Revisited

Q-mode Eigenvector Analysis: Selected Figures

13. Group Distance and Separation

14. Path Analysis and Structural Equation Modeling

Path Analysis: selected figures

15. Linear models with categorical data

16. Introduction to Likelihood

DOCUMENTS FOR PROGRAMMING IN R

NB: Also see bits of code in lecture notes (for Q-mode similarity; principal coordinates, canonical variates, canonical correlations...)

Foote's tutorial:

R tutorial Autumn 2011

Gene Hunt's tutorial:

Hunt-1. R basics, introduction to statistics and plotting

Hunt-2. Importing data, statistical models in R, biodiversity analyses

Hunt-3. Data manipulation, multivariate analysis

Hunt-4. Programming: loops, functions, etc.

Hunt-5. Phylogenetic comparative methods, re-sampling procedures

Miscellaneous scripts

R routine for standarization and/or mean-centering of variables

R routine for row normalization

R routine for calculating cosine theta similarity between rows

R routine for Q-mode cluster analysis

R routine for R-mode cluster analysis

R routine for Q-mode non-metric multidimensional scaling

R routine for R-mode non-metric multidimensional scaling

R routine for R-mode principal component analysis

R routine for singular value decomposition

R routine to reproduce Bookstein et al.'s factor analysis of Wright's leghorn data