Book contents
- Frontmatter
- Contents
- Preface
- 1 Basic Concepts in Probability and Statistics
- 2 Hypothesis Tests
- 3 Confidence Intervals
- 4 Statistical Tests Based on Ranks
- 5 Introduction to Stochastic Processes
- 6 The Power Spectrum
- 7 Introduction to Multivariate Methods
- 8 Linear Regression: Least Squares Estimation
- 9 Linear Regression: Inference
- 10 Model Selection
- 11 Screening: A Pitfall in Statistics
- 12 Principal Component Analysis
- 13 Field Significance
- 14 Multivariate Linear Regression
- 15 Canonical Correlation Analysis
- 16 Covariance Discriminant Analysis
- 17 Analysis of Variance and Predictability
- 18 Predictable Component Analysis
- 19 Extreme Value Theory
- 20 Data Assimilation
- 21 Ensemble Square Root Filters
- Appendix
- References
- Index
15 - Canonical Correlation Analysis
Published online by Cambridge University Press: 03 February 2022
- Frontmatter
- Contents
- Preface
- 1 Basic Concepts in Probability and Statistics
- 2 Hypothesis Tests
- 3 Confidence Intervals
- 4 Statistical Tests Based on Ranks
- 5 Introduction to Stochastic Processes
- 6 The Power Spectrum
- 7 Introduction to Multivariate Methods
- 8 Linear Regression: Least Squares Estimation
- 9 Linear Regression: Inference
- 10 Model Selection
- 11 Screening: A Pitfall in Statistics
- 12 Principal Component Analysis
- 13 Field Significance
- 14 Multivariate Linear Regression
- 15 Canonical Correlation Analysis
- 16 Covariance Discriminant Analysis
- 17 Analysis of Variance and Predictability
- 18 Predictable Component Analysis
- 19 Extreme Value Theory
- 20 Data Assimilation
- 21 Ensemble Square Root Filters
- Appendix
- References
- Index
Summary
The correlation coefficient measures the linear relation between scalar X and scalar Y. How can the linear relation between vector X and vector Y be measured?Canonical Correlation Analysis (CCA) provides a way. CCA finds a linear combination of X, and a (separate) linear combination of Y, that maximizes the correlation. The resulting maximized correlation is called a canonical correlation. More generally, CCA decomposes two sets of variables into an ordered sequence of component pairs ordered such that the first pair has maximum correlation, the second has maximum correlation subject to being uncorrelated with the first, and so on. The entire decomposition can be derived from a Singular Value Decomposition of a suitable matrix. If the dimension of the X and Y vectors is too large, overfitting becomes a problem. In this case, CCA often is computed using a few principal components of X and Y. The criterion for selecting the number of principal components is not standard. The Mutual Information Criterion (MIC) introduced in Chapter 14 is used in this chapter.
Keywords
- Type
- Chapter
- Information
- Statistical Methods for Climate Scientists , pp. 335 - 365Publisher: Cambridge University PressPrint publication year: 2022