12.864 -- Inference from Data and Models
Course Description: Fundamental methods used for exploring the information content of observations related to kinematical and dynamical models. Basic statistics and linear algebra for inverse methods including singular value decompositions, control theory, sequential estimation (Kalman filters and smoothing algorithms), adjoint/Pontryagin principle methods, model testing, etc. Second part focuses on stationary processes, including Fourier methods, z-transforms, sampling theorems, spectra including multi-taper methods, coherences, filtering, etc. Directed at the quantitative combinations of models, with realistic, i.e. sparse and noisy observations.

This class is at the Graduate level
Instructor: C. Wunsch
Open courseware website: http://ocw.mit.edu/OcwWeb/Earth--Atmospheric--and-Planetary-Sciences/12-864Spring-2005/CourseHome/in
Prerequisites: 18.075 or 18.086

Insider's Wisdom

This class is specially designed for students that work in Earth, Atmospheric and Planetary Sciences, but can be usuful for anyone using large datasets and often models that are based on fluid dynamics or other differential models.

The class includes time series, data simulation and state estimation. The goal is for the students to walk away with a toolkit that they can use in the earth sciences, and that they can read the academic literature in the field.

This course is at the Graduate level, but undergraduates have succesfully completed the course.

Back to Classes