This is an overview of practical machine learning.
There are a large number of machine learning
classes out there, and they are often very good.
So, the focus of this class will be
primarily on hand-drawn application of machine learning in R.
So, the idea being that we'll try to focus on the R packages and the
ideas that will allow you to actually take
data and perform machine learning on those data.
We'll also talk a little bit conceptually about each of these prediction
methods work and maybe some of the cases where there might be trouble.
And then we'll point you to resources where you can learn more in depth about
the mathematical details or the really hardcore
computational underpaying of these methodologies if you're interested.
So the Practical Machine Learning Content.
We'll start with prediction study design, we'll talk about cross validation.
The caret package for prediction in R, some pre-processing.
Predicting what the variety of different ideas like regression and trees.
We'll talk about common ideas like boosting, bagging,
model blending, and a little bit about forecasting.
So here are some examples of things we'll cover.
We'll cover basic terms, like what are true positives and false positives?
What are true negatives and false negatives,
sensitivity and specificity, those sorts of things.
We'll also cover how to deal with correlated
predictors by preprocessing out data that had correlated predictors.
When we're moving them from the training data
set, and we'll talk a little bit about boosting.
So, this is a very more technical machine learning idea, but can be applied
quite simply using the functions of a R to really improve your prediction accuracy.