Quick Links

Data Query
Universal Query Interface
^experimental

Related Sites

21 Nov 2009 23:40 UTC 2009325+2340 UTC

Recent Changes Page History Upload Edit Page Print View

CMSS 6352: Environmental Forecasting

Spring 2009

Course Description:

Statistical techniques (classic and Bayesian) and new artificial intelligence based techniques, such as neural networks, for the analysis of and forecasting of environmental systems. Emphasis on the forecasting of coastal and marine systems. Prerequisites: CMSS 6303, Applied Multivariate Statistical Analysis, CMSS 6405, Natural System Modeling or approval of instructor.

Course Progression
Class	Topic Covered	Reading	To Do List
W - Jan 14	No class - class start on Friday Jan 16	NA	NA
F - Jan 16	Class introduction, Discussion of environmental systems	NA	NA
M - Jan 19	Martin Luther King Day - No class	NA	NA
W - Jan 21	Different types of systems and modeling techniques, chaotic nature of atmosphere	NA	NA
F - Jan 23	Invited speaker - Wind energy	NA	NA
M - Jan 26	Modeling techniques	Read Wilks Chapters 1 & 2
W - Jan 28	Modeling Techniques & Review of probability and statistical concepts	Read Trauth Chapts 1 & 2	Select computational environment and implement Trauth Chapts 1 & 2
F - Jan 30	Review of probability and statistical concepts
M - Feb 2 2hrs	Review of probability and statistical concepts	Read Wilks Chapt. 3.1, 3.2	Hmwrk #1: find time series, assess quality (%data available, max gap length,...) plot time series and histogram, compute mean, median, range, standard deviation using computational env.
W - Feb 4 2hrs	Review of probability and statistical concepts	Read Wilks Chapt. 3.3, 3.4
F - Feb 6	No class
M - Feb 9	No class
W - Feb 11	No class
F - Feb 13 1hr	Review of probability and statistical concepts	Read Wilks Chapt. 3.5, 3.6	Hmwork #2: Wilks Problems 2.3, 2.4, 2.5. Divide your environmental time series in time intervals (months for example) or in categories, plot the average for each interval/category with that interval/category variance, discuss if there seems to be a trend (no need for statistical assessment at this point), draw a scatter plot between your time series and one or more likely forcing, suggest two methods to forecast the quantity (no details at this point)
M - Feb 16 2hrs	End Rev. probability and statistical concepts
W - Feb 18 2hrs	Data Imputation - Wilks 4-5
M - Feb 23	No class
W - Feb 25 2hrs	Wilks 4-5 & Statistical Forecasting	Read Wilks 4.4.5, 4.5.2, 5.2.5, 6.1-6.2	Hmwork #3:Discretization Homework
F - Feb 27 1hr	HRI Seminar: Ken Burnham	The fundamental ideas of AIC-based model selection and multimodel inference
M - Mar 2 1.5hrs	Artificial Neural Networks
W - Mar 4 1.5hrs	Artificial Neural Networks		Hmwork #4: On individual pages plot your time series, a Q-Q plot to assess normality of your time series (comment on the plot, i.e. what type of distribution as compared to normal distribution), plot the autocorrelation function for your main variable and comment on predictive possibilities (see course Matlab code or use other tool)
M - Mar 9 1.5hrs	ANN Training
W - Mar 11 1.5hrs	ANN Applications to class data		Hmwork #5: Multi linear analysis and Predictor selection. Select 1 predictand & 3 predictors (at least 2 different variables) and proceed with a forward and a backward process to select a "good" multi linear regression model. Add brief comments to the selection process.
Mar 16-20	Spring Break
M - Mar 23 1.5hrs	ANN & Statistics Models	Read Marzban (2)
W - Mar 25 1.5hrs	ANN / Cycles & Periodograms	Read Marzban (2), Wilks 8.4
M - Mar 30	No Class
W - Apr 1 1.5hrs	Cycles & Periodograms	Read Wilks 8.5	Proposals Due
M - Apr 6 1.5hrs	Verification	Read Wilks 7/Marzban (3)	Hmwrk #6: Include periodogram in project
W - Apr 8 1.5hrs	Verification	Read Wilks 7/Marzban (3)
M - Apr 13 1.5hrs	Statistical Forecasting	Read Wilks 6	Proposal Reviews Due
W - Apr 15 1.5hrs	Statistical Forecasting	Read Wilks 6
M - Apr 20 1.5hrs	More Time Series	Read Wilks 8.3
W - Apr 22 1.5hrs	More Time Series	Read Wilks 8.3
M - Apr 27 1.5hrs	Random Forests
W - Apr 29	No Class
F - May 1	Visit of CCWFO 10AM-12PM
M - May 4 2.5hrs	Project Presentations Class Starts at 10AM		Project Presentations
W - May 6	Reading Day
T - May 12			Project Reports Due
F - May 15			Take Home Exams Due