Exam SRM Details

Statistical learning helps us study data to find answers and solve problems. You'll learn the difference between supervised and unsupervised learning, the difference between regression and classification, and how to assess model accuracy. You'll also learn basic ways to examine data numerically and graphically.

We will begin our journey of statistical learning by looking at models where the response variable is normally distributed. We will focus on using ordinary least squares as our estimator in linear regression, and build upon this fundamental concept. Then, we will diagnose these models and determine if they are appropriate for the given dataset. Lastly, we will see how other similar, but different approaches compare to an ordinary least squares model.

The next topic is generalized linear models, where the response variable can be modeled with a distribution other than normal. This involves learning about the linear exponential family of distributions and the use of maximum likelihood estimation. We will examine the similarities and differences compared to linear regression. In addition, we take a closer look at the results from modeling the response variable with Bernoulli and Poisson distributions.

In this section, we will explore time series models and other important concepts related to time series, such as stationarity and autoregression. We will also learn how to model time series and forecast future observations.

Most supervised learning methods discussed so far are parametric methods. Now, let's venture into decision trees, which are non-parametric. We will observe how decision trees work in both regression and classification settings, and explore other statistical concepts that can be applied to trees.

You've learned a lot about supervised learning. Time to shift gears to learn about unsupervised learning. This section will cover two unsupervised learning methods, principal components analysis (PCA) and clustering analysis.