The Bureau of Labor Statistics cites statistics as one of the fastest-growing fields in the U.S., with a projected growth rate of 31% between 2018 and 2028. So, individuals with strong statistical analysis skills are in high-demand. Beyond this, statistical jobs are cited as having relatively low stress and great pay. Statistician positions top multiple job lists online, including
This course delves deeper into ideas of linear regression. We move beyond the simple linear regression model -- a single quantitative variable used to predict a quantitative response -- and allow multiple quantitative and categorical predictors. After a short review of simple linear regression, multiple linear regression models are formulated with a general matrix-vector notation. Applications, analyses, and interpretations of results are a focus in this course. We will use R throughout the term.
Multiple linear regression is a key analysis tool for statisticians and engineers. Our approach to this topic begins by deriving least squares estimators from the general matrix-vector format of the linear model. Then, we explore topics associated with linear regression, including assessment of assumptions for statistical inference, formal statistical inference, model fit, predictions, multi-collinearity, and variable selection. Specific models, including polynomial regression, categorical interactions, and poisson regression, are covered. We will also provide statistical insights to the LASSO and elastic net models you may have seen previously in a machine learning course.
This class is scheduled in a hybrid format. Specifically, we will all meet in a traditional classroom setting twice a week: Mondays and Thursdays. The face-to-face class meetings will have a similar feel to a traditional lecture (e.g. instructor-led content and activities). Thus, attendance of class on those days is a requirement. Tuesdays and Fridays are online/asynchronous class days. The online days will contain a mixture of new content, data analysis examples using statistical software, and structured time for student groups to work on class assignments.
Although we will not meet at a specific time on Tuesday and Friday, it is expected that students work through the provided material at some point during the specified day. The face-to-face meetings will operate under the assumption that the previous days' online material was completed. Thus, for you to be successful in MA 485, it is extremely important to stay on schedule.