Course
Announcement for SMS 595
Dr. Huijie Xue
University of Maine
Phone: 581-4318
Provides theoretical
and computational guidance on techniques commonly used in the data analysis.
The first half of the course will cover regression methods, and the second half
will cover time series analysis and digital filters. Real data will be used to
illustrate the practical aspects of the subject with emphasis on developing a
hands-on understanding of the methods and correct interpretation of results.
Prerequisites: MAT 127 or equivalent. Cr. 3.
Syllabus
|
Week
#1 |
Review
of Calculus and Statistics |
-
Functions and Derivatives, -
Maxima, and Mimima -
Expectation, (Co)Variance, Statistical tests |
|
Week
#2 |
Linear
Regressions |
-
Type I linear regression -
Error Analysis -
Correlation |
|
Week
#3 |
|
-
Rank Correlation -
Type II Linear Regressions |
|
Week
#4 |
Other Regression Methods |
-
Linear Multiple Regression -
Polynomial Regression -
Variance Analysis -
Nonlinear Regression |
|
Week
#5 |
Time Series Analysis |
-
Sampling and Aliasing -
Time Domain Techniques |
|
Week
#6 |
|
-
Harmonic Analysis |
|
Week
#7 |
Review and Mid-Term Exam |
|
|
Week
#8 |
Non-Parametric Spectral Analysis |
-
Fourier Transform -
Autocovariance Method -
Periodogram |
|
Week
#9 |
|
-
Spectral Leakage -
Data Tapers |
|
Week
#10 |
|
-
Band- and Block Averaging -
Bandwidth -
Confidence interval |
|
Week
#11 |
Parametric
Spectral Analysis |
- Auto
Regressive Processes -
Parametric Spectral Analysis -
Maximum Entropy Method |
|
Week
#12 |
Cross-Spectral Analysis |
-
Cross Spectra -
Rotary Spectra -
Rotary Cross-Spectra |
|
Week
#13 |
Digital
Filters |
-
Ideal Filters -
Design Digital Filters -
Commonly Used Filters -
Recursive Filters |
|
Week
#14 |
EOF
Analysis |
- Mathematical
Procedure -
Examples -
Interpretations |
|
|
Final Exam (Project and
Presentation) |
|
Text(s):
-
Edward Laws
(1997): Mathematical Methods for Oceanographers. John Wiley & Sons, Inc.
- W.J. Emery and R.E. Thomson (2001): Data Analysis Methods in Physical Oceanography, ELSEVIER