EOSC 510 (Data Analysis in Atmospheric, Earth and Ocean Sciences)

Course description and outline.    Textbook (optional)

Week 1 (Sept. 3-8):

Chapter 0 (Introduction and course setup)

Ch0.mp4 (video file, duration 18 min., size 38 MB),  Ch0.pdf (pdf file of the slides used in the video lecture)

Chapter 1. Correlation and regression

Ch1a.mp4 (20 min., 18 MB) (Mean and variance),   Ch1.pdf,   Ch1_Q_solns.pdf (pdf file containing solutions to questions posed in the videos)
Ch1b.mp4 (26 min., 25 MB) (Correlation)
Ch1c.mp4 (16 min., 14 MB) (Linear regression)
Ch1d.mp4 (21 min., 20 MB) (Multiple linear regression)
Ch1matlab.mp4 (18 min., 30 MB) (Matlab programming)

Week 2 (Sept. 9-15):

Chapter 2. Principal component analysis (PCA) and rotated PCA

Ch2a1.mp4 (16 min., 15 MB) (Geometric approach; eigenvector approach),   Ch2a.pdf,   Ch2_Q_solns.pdf
Ch2a2.mp4 (24 min., 22 MB) (Eigenvector approach, continued)
Ch2a3.mp4 (19 min., 18 MB) (Complex data; orthogonality)
Ch2a4.mp4 (22 min., 22 MB) (PCA applied to real data)

Ch2b1.mp4 (18 min., 16 MB) (Scaling; degeneracy),   Ch2b.pdf
Ch2b2.mp4 (18 min., 16 MB) (Smaller covariance matrix; mean removal)

Week 3 (Sept. 16-22):

Ch2b3.mp4 (16 min., 15 MB) (Singular value decomposition)
Ch2b4.mp4 (17 min., 16 MB) (Missing data; significance tests)

Ch2c1.mp4 (22 min., 24 MB) (Rotated PCA),   Ch2c.pdf
Ch2c2.mp4 (24 min., 25 MB) (Varimax; teleconnection patterns)
Ch2c3.mp4 (17 min., 16 MB) (PCA versus Rotated PCA)
Ch2c4.mp4 (20 min., 18 MB) (PCA for vectors)

Week 4 (Sept.23-29):

Chapter 3. Canonical correlation analysis (CCA)

Ch3a.mp4 (25 min., 23 MB) (CCA theory),   Ch3.pdf,   Ch3_Q_solns.pdf
Ch3b.mp4 (15 min., 12 MB) (CCA theory, continued)
Ch3c.mp4 (22 min., 20 MB) (Pre-filter by PCA)
Ch3d.mp4 (14 min., 12 MB) (Maximum covariance analysis)

Chapter 4. Time series

Ch4a1.mp4 (19 min., 16 MB) (Fourier spectral analysis: autospectrum),   Ch4a.pdf,   Ch4_Q_solns.pdf
Ch4a2.mp4 (13 min., 11 MB) (Autospectrum, continued)
Ch4a3.mp4 (14 min., 13 MB) (Autospectrum, continued)

Week 5 (Sept.30-Oct.6):

Ch4a4.mp4 (20 min., 18 MB) (Cross-spectrum)


Ch4b1.mp4 (14 min., 12 MB) (Windows),   Ch4b.pdf.
Ch4b2.mp4 (14 min., 13 MB) (Filters)
Ch4b3.mp4 (20 min., 19 MB) (Filters, continued)

Ch4c1.mp4 (22 min., 21 MB) (Singular spectrum analysis),   Ch4c.pdf.
Ch4c2.mp4 (17 min., 18 MB) (Multichannel singular spectrum analysis)

Week 6 (Oct.7-Oct.13):

Chapter 5. Classification and clustering

Ch5a1.mp4 (20 min., 18 MB) (Classification: k-nearest neighbour classifier),   Ch5a.pdf,   Ch5_Q_solns.pdf
Ch5a2.mp4 (15 min., 12 MB) (Conditional probabilities)
Ch5a3.mp4 (17 min., 16 MB) (Bayes' theorem)
Ch5a4.mp4 (21 min., 19 MB) (Logistic regression)


Ch5b1.mp4 (14 min., 13 MB) (Clustering: k-means clustering),   Ch5b.pdf.
Ch5b2.mp4 (25 min., 24 MB) (Hierarchical clustering)


Ch5c.mp4 (28 min., 25 MB) (Self-organizing maps),   Ch5c.pdf.

Week 7 (Oct.14-Oct.20):

Chapter 6. Feed-forward neural network models

Ch6a1.mp4 (12 min., 11 MB) (McCulloch and Pitts model),   Ch6a.pdf,   Ch6_Q_solns.pdf
Ch6a2.mp4 (14 min., 13 MB) (Perceptrons)
Ch6a3.mp4 (13 min., 11 MB) (Limitations of perceptrons)
Ch6a4.mp4 (15 min., 14 MB) (Multi-layer perceptrons, MLP)
Ch6a5.mp4 (23 min., 23 MB) (MLP, continued)

Week 8 (Oct.21-Oct.27):

Ch6a6.mp4 (13 min., 13 MB) (MLP, continued)
Ch6a7.mp4 (23 min., 23 MB) (Back-propagation)
Ch6a8.mp4 (19 min., 19 MB) (Hidden neurons)

Midterm exam: Tuesday, 22 Oct. 11:00-12:15 (Pacific time)

Week 9 (Oct.28-Nov.3):

Ch6b1.mp4 (20 min., 19 MB) (MLP classifier),   Ch6b.pdf
Ch6b2.mp4 (29 min., 28 MB) (Radial basis functions, RBF)
Ch6b3.mp4 (19 min., 18 MB) (Conditional probability distributions)
Ch6b4.mp4 (17 min., 16 MB) (Mixture models)

Chapter 7. Nonlinear optimization

Ch7a.mp4 (19 min., 18 MB) (Nonlinear optimization),   Ch7.pdf,   Ch7_Q_solns.pdf
Ch7b.mp4 (21 min., 19 MB) (Gradient descent methods)

Week 10 (Nov.4-10):

Ch7c.mp4 (16 min., 17 MB) (Evolutionary computation and genetic algorithms)
Ch7d.mp4 (14 min., 22 MB) (Evolutionary computation: robotic soccer)
Optional youtube videos on genetic algorithms: GA learns to fight (2 min.), GA fish (1 min.)

Chapter 8. Learning and generalization

Ch8a1.mp4 (14 min., 15 MB) (Mean squared error and maximum likelihood),   Ch8a.pdf,   Ch8_Q_solns.pdf
Ch8a2.mp4 (31 min., 29 MB) (Objective functions and robustness)
Ch8a3.mp4 (14 min., 13 MB) (Variance and bias errors)
Ch8a4.mp4 (21 min., 21 MB) (Regularization)
Ch8a5.mp4 (17 min., 19 MB) (Cross-validation)

Week 11 (Nov.11-17):

Ch8b1.mp4 (17 min., 18 MB) (Bayesian neural networks; ensemble of models),   Ch8b.pdf
Ch8b2.mp4 (13 min., 12 MB) (Errors of ensembles)
Ch8b3.mp4 (21 min., 20 MB) (Errors of ensembles, continued)
Ch8b4.mp4 (13 min., 13 MB) (Nonlinear ensemble averaging; boosting)
Ch8b5.mp4 (10 min., 10 MB) (Linearization from time-averaging)
Ch8b6.mp4 (9 min., 9 MB) (Regularization of linear models: ridge regression & lasso)

Chapter 9. Tree-based methods

Ch9a.mp4 (13 min., 12 MB) (Classication and regression trees (CART)),   Ch9.pdf,   Ch9_Q_solns.pdf

Week 12 (Nov.18-24):

Ch9b.mp4 (15 min., 15 MB) (CART applied to ozone in Los Angeles)
Ch9c.mp4 (17 min., 17 MB) (pruning trees; CART for clasification)
Ch9d.mp4 (10 min., 10 MB) (Random forests, boosted trees)

Chapter 10. Forecast verification

Ch10a.mp4 (12 min., 12 MB) (Binary forecasts),   Ch10.pdf,   Ch10_Q_solns.pdf
Ch10b.mp4 (18 min., 17 MB) (Binary forecasts, continued)
Ch10c.mp4 (20 min., 21 MB) (Skill scores)
Ch10d.mp4 (13 min., 12 MB) (Probabilistic forecasts)

Week 13 (Nov.25-Dec.1):

Chapter 11. Nonlinear principal component analysis (NLPCA)

Ch11a.mp4 (13 min., 14 MB) (Auto-associative neural networks for NLPCA; open curves),   Ch11.pdf
Ch11b.mp4 (21 min., 21 MB) (NLPCA of sea surface temperature anomalies)
Ch11c.mp4 (15 min., 15 MB) (Closed curves)

Chapter 12. Kernel methods (optional material)

Ch12a.mp4 (12 min., 12 MB) (From neural networks to kernel methods),   Ch12.pdf
Ch12b.mp4 (17 min., 15 MB) (Advantages and disadvantages of kernel methods)
Ch12c.mp4 (20 min., 19 MB) (Support vector regression)

Summary

summary.mp4 (11 min., 15 MB),   summary.pdf

Final exam: Friday, 6 Dec. 12:00-14:30 (Pacific time)