SOURCE 2019

​Watch this session live or the recording later.​​​

Brain Hooper (Graduate Student), Heather McKinnon (Graduate Student), and Divya Chadrika Kalla (Graduate Student) - Annuity Profit Simulation
We will present an R script for simulating a large number of whole life single annuity premiums prices for an insurance company, and to produce projected annual profits for the company. The program uses mortality data from the Society of Actuaries to create a life table which will calculate the annuity expected present value for given ages. The program accepts user input of a starting age range, maturity age, monthly annuity benefit, interest rate, number of policy holder lifetimes to generate, number of company years to simulate, and an ROI interest rate. Using the life table data and user input, the program creates a simulated number of policy holders’ lifetimes with random starting and death ages and their calculated premiums. This program also simulates a profit for a projected number of years to measure the insurance company’s annual profit. Several graphs and tables are produced to help illustrate these calculations. The calculations and design of this program were advised by Dr. Chin-Mei Chueh, actuarial science professor and Society of Actuaries council member, and Dr. Donald Davendra, computer science professor.

Brian Hooper (Graduate Student) - Emotion classification
We present an emotion prediction system that classifies electroencephalography brain activity data into one of four emotion categories. Emotion classification is inherently difficult because of the subjective nature of emotions, thus our emotion model uses two-dimensional values of valence and arousal for classifying an individual emotional state. The EEG data was provided from the DEAP dataset, containing 40-channel EEG data from 32 participants who each watched 40, one-minute long excerpts of music videos and labeled their emotional states during each video. We demonstrate the unique challenges with working with EEG data, our methods for dimension reduction and classification, and the results we obtained using our classification model.

Aliyah Pana - Compartmental Models for Infectious Disease
Mathematical modeling is the process of using various mathematical structures to represent real world situations. These models can be used to predict pandemics, natural disasters, population data, and other real world aspects. We can create a model to track a disease’s possible spread. By applying real world data, we can create a simulation of a disease. Using the information gathered from the model, we can understand how an outbreak may behave. The results can be put into perspective to create effective precautions and actions to combat an outbreak.

Rachel Walker, Olivia Vasquez, and Riley Krall - An Optimized Drone Fleet Response System
We present an optimized drone fleet disaster response system modeled after the recent hurricane in Puerto Rico. Our model uses the assumptions given by HELP Inc., a non-profit organization, to efficiently deliver supplementary medical supplies to hospitals in effected areas. We give an algorithm for determining optimal drop locations for multiple ISO cargo containers, contained in which are drones and medical supplies. We also give an algorithm for selecting a drone fleet from predefined options and optimal packing configurations for ISO cargo containers. Finally, we apply our model to Puerto Rico detailing efficient drone routes and schedules, and conclude with a sensitivity analysis of our model.