MATH5055

The course covers common predictive models used by actuarial and financial practitioners. Emphasis is on concepts and applications while theory is left for more advanced courses. Topics include Exploratory Data Analysis, regression/classification, supervised/unsupervised learning, Cross Validation, model tuning, GAMs, GLMs, LASSO, KNN, Decision Trees, Random Forests, GBM, and feed forward Neural Networks. Local industry practitioners will present real-world applications over the final four weeks of the course. The software R is used throughout. We currently expect Optum, Travelers, Thrivent, and Securian to present during the last
four weeks.

prereq:
STAT 3032, Regression and Correlated Data is the only formal perquisite.
However, the important things is that the students are already familiar with:
1. Basic Statistics and Probability
2. Linear Regression
3. Using R
Students who meet the latter may, at the instructor’s discretion, be admitted to the course.