Welcome to this course introduction on
Coursera on Computational Methods for Data
Analysis. My name is Nathan Kutz. I am a
professor and chair of applied mathematics
at the University of Washington. I'd like
to kind of briefly introduce the ideas of
what I want to cover in this course so
that you would have a better idea of, of
how to think about this course. Data, of
course, has become ubiquitous in all the
sciences, as well as things that are not
necessarily considered sciences, such as
computer science where you're thinking
about data mining applications. But in
either case, data is driving revolutions
in how we think about managing and
analyzing both complex dynamical systems
as well as things that are purely data
driven, such as market place factors as
well as thinking about mining for data
bases for important information. This
course has a very specific aim which is to
integrate many aspects of data analysis in
a very broad sense. In the set we're going
to start off with, in fact with signal
processing and dynamics of signal
processing. We're gonna look at thinking
about statistical methods, bringing them
into these ideas of how to think about
time and frequency analysis, and building
this forward towards understanding things
like dimensionality reduction, and
principal component analysis. Also
included in this is going to be some kind
of idea of thinking about machine
learning. How do you think about
integrating dynamics? How do you integrate
your data collection into teaching a
computer to see bigger trends in your
data? And in constructing appropriate
models that might be able to help you
predict future states of your system of
interest. This course is an upper division
undergraduate, beginning graduate course,
so it's fairly high-level. It's expected
that you'd have an amazing intro,
background in linear algebra. Well,
amazing may be a little oversh, pushing
but you should have a very good background
in linear algebra. You should have a solid
background in some computing techniques.
here, the course wi ll be mostly based
upon Matlab. It's a very fun course,
there's lots of great and relevant
examples. it will also make you work very
hard. so have no illusions about this
being easy. It is a very difficult course,
but at the end, you'll learn some of those
cutting edge techniques being used today
in both thinking about data mining
applications, but also how you apply this
to applications where you might have an
underlying physical system of interest and
you want to use the data from that system
to help you be able to say something very
concrete about your system of interest by
using data from that system directly in
meaningful ways. I hope you enjoy the
course, and welcome.