STA 294 (b): SPATIAL STATISTICS - Dave Higdon, Spring 1998

Coverage: Theory and practice of applying models to spatial data. The course will cover basic theoretical concepts as well as practical implementation issues. We'll look at examples from image analysis, biology, ecology, and genetics, just to name a few.

Part 1 Introduction to Gaussian distributions and Gaussian processes. Spatial covariance structures, simulation, conditional distributions and standard Geostatistical/Kriging methodology.

Part 2 Lattice models, non Gaussian spatial models, and other alternatives to standard Geostatistics.

Part 3 (if time permits) Spatial point process models. The S+Spatial Statistics Software will Support the Course.

Recommended Reading:

S+ Spatial Statistics Manual.

Cressie, N.A.C. (1991) Statistics for Spatial Data

Ripley, B.D. (1981) Spatial Statistics

Isaaks, E.H. and Srivastava, R. M. (1990) An Introduction to Applied Geostatistics

Classes are Mondays 2:20 pm-5:10 pm in Old Chem 025,

The wavelet part [STA294 a] is on Jan. 19, 26, Feb 9,23, Mar 2, 30, and Apr 13!

The spatial statistics part [STA294 b] is on Jan. 19, Feb 2, 16, Mar 9,23, and Apr 6, 20!

Day	Topics
Jan. 19	Introduction
Feb. 2	Gaussian Distributions and processes. Spatial covriance models.
Feb. 16	Geostatistics/Kriging. Interpolation; estimation; block Kriging and support; Bayesian Kriging.
Mar. 9	Model building; Model checking; simulation.
Mar. 23	Lattice models; Image models; Markov random fields.
Apr. 6	later
Apr. 20	later

Prerequisites: Consent of instructor. Familiarity with programming in C, FORTRAN, SPlus, or SAS is desirable.

Please send comments to higdon@stat.duke.edu