Academics / Courses / DescriptionsIEMS 395-490: Special Topics in IE: Applied Statistical Learning and Decision Making
Academics
/ Courses
/ Descriptions
VIEW ALL COURSE TIMES AND SESSIONS
Prerequisites
IEMS 304 or CS 349 or equivalentDescription
Description
- Will explore common problems related to finance and healthcare and how to tackle them using statistical tools. Common techniques we will explore include poisson-regression, binning, bagging (credit rating), time-series analysis methods like unit-root test (stock market), Tweedie regression, and survival analysis (healthcare).
- Bi-weekly homework, midterm, final exam, class project. Lab attendance mandatory.
- This special topics course can be used as an IE/OR elective for Industrial Engineering.
LEARNING OBJECTIVES
- Familiarizing students with some of the recurring statistical questions in finance and healthcare industry.
- Designing, conducting, and predicting loan recovery, patient mortality etc. from messy incomplete datasets that are prevalent in the real world.
- Making informed decisions with confidence. Statistics will not indicate whether a loan should be given. The practitioner will have to decide based on factors.
- Emphasizing Ex-post over Ex-ante. Providing depth of understanding about a dataset and then choosing a statistical tool to tackle that problem, rather than applying a method on a dataset and improving it based on the outcome.
TOPICS
There are two broad topics of focus:
- Making important decisions about financial risk management
a. Credit scoring: Score creditworthiness
b. Risk Modelling: Days-past-due and loan recovery predictions
c. Trend analysis
d. (optional) Fraud detection: Predicting willful defaults
- Making important decisions about healthcare
a. Dose escalation: Finding Toxicity probability of a dose.
b. Survival Analysis: Estimating the time until a specific event (e.g., death, relapse)
c. (optional) QTL mapping.
- Project (tentative): The students will follow the trails of what led to the 2008 financial crisis, and how the ACA helped dampen its effects in healthcare.
MATERIALS
No-required textbook. Optional reading will be presented in class. Computational software: R/Rstudio