2012 AMRE Annual Report

Modeling Employee Flight Risk for American International Group, Inc.

Student Researchers: Ana Godonoga, Matthew Lambert and Allie Webb,
Faculty Advisors: Lisa Verdon (Business Economics) and John Ramsay (Mathematics)

The goal of this project was to develop a model that will allow AIG to predict, based on key variables, an employee’s likelihood of leaving the company. The work toward this goal resulted in a variety of deliverables provided for AIG. First, the AMRE team cleaned and recoded several databases of uncombed data which were then merged into one master data file to be used for statistical analysis. In addition to this new “statistics-friendly” database, a data cleaning program was created in Microsoft Excel. This program gives AIG a tool that will speed the data cleaning process as they add new data for analysis. Third, some static snapshot analysis was performed on the new master data file. Key variables in flight risk prediction were determined through literature research and statistical tests. A few noteworthy findings were shared with AIG with recommendation to pursue this analysis further as more data becomes available. Finally, a mathematical model for predicting employee flight risk was created. The model is based on Probit analysis and assigns a flight risk probability to each employee based on key profile variables. AIG was provided a final set of recommendations for ways to improve their systems and data gathering techniques as they continue with their predictive analytics efforts.

Assessment Framework for Experiential Learning at The College of Wooster

Student Researchers: Julia Land, Giang Nguyen, and Eric Petry
Faculty Advisors: Pam Pierce (Mathematics) and Lisa Verdon (Business Economics)

This project is a continuation of the 2011 AMRE project: College of Wooster Assessment. We conducted research on experiential learning (EL) in general and EL opportunities at Wooster in particular. Our goals were: 1) create clear guidelines for each tier, along which EL programs can be classified and 2) develop assessment tools for EL programs which reflect the unique nature of each program. We investigated and developed a model for EL, which is the foundation for our three-tiered classification framework. In the assessment component, we created assessment tools and assessment processes pertinent to two different groups of EL programs. These tools should help the College make more informed and effective decisions regarding the allocation of support to and investment in EL programs. We also outlined the EL process that students might have to undertake in order to participate in EL experiences.

Modeling Tension in Wound Polyester Film

Student Researchers: Kevin Dinh and Sarah Laper
Faculty Advisors: Drew Pasteur (Mathematics) and Susan Lehman (Physics)

This project applied winding concepts such as radial pressure, yield point, Young’s modulus, friction, and Hakiel’s simple winding model to predict the occurrence of telescoping and dimpling, two common winding defects at Kent Displays©, in polyethylene terephthalate (PET) rolls, which are used to make Boogie Boards. Hakiel’s simple winding model was closely examined and implemented in MATLAB® during the early stages of this project. After validating the predicted pressure measurements from Hakiel’s model with measurements taken at Kent Displays©, telescoping and dimpling were then examined. In addition to discovering the location and frequency of telescoping and dimpling in PET rolls, common winding variables, such as taper percent;roll length; initial tension; radial, tangential, and core modulus; were closely studied. Lastly, a list was compiled on how altering winding variables affects radial pressure in a wound roll.

Undergraduate Research in Knot Theory

Student Researchers: Danielle Shepherd, Joseph Smith, and Sarah Smith-Polderman
Faculty Advisors: Jennifer Bowen and John Ramsay (Mathematics)

This project is a continuation of past research. During our time this summer, our team created a digital catalogue of over 50 Klein links. In addition, we investigated connections between the invariants of torus links and Klein links with the goal of further understanding the effects of our construction. Through this research, we were able to discover results that related the crossing number of torus links to Klein links and Klein links to Klein links. After developing these ideas further, we created a paper that involved these findings and hope to publish these results in the future. We will be presenting a portion of our findings at the 2012 UnKnot Conference at Denison University later in the summer.

Molecular Imaging Using Kinect

Student Researchers: Norman Chamusah and Benn Snyder
Faculty Advisors: Denise Byrnes (CS) and Nicholas Shaw (Chemistry)

The goal of this project is to extend the work of the Molecular Playground for use in an educational environment. The Molecular Playground is a hardware-software installation that allows the user to rotate 3-D chemical molecules. This is done by capturing hand positions detected by the Kinect camera and sending these positions to a 3-D modeler. Our extensions also support 3-D rotations and add x-y translations, molecule selection and partial molecule rotation. This allows the chemistry instructor to interact more naturally with the model and without direct contact with the hardware.

Finding regulatory motifs in the Solanum Lycopersicum genes

Student Researcher: Michelle Blackwood
Faculty Advisor: Sofia Visa (Computer Science)

(research performed at Ohio Agricultural Research and Development Center-OARDC)

Transcription factor binding sites (motifs) are small portions of proteins that are homologous to regions in other proteins. To date, such motifs of genes in the Solanum Lycopersicum (tomato) have not been discovered. This research seeks to find potential transcription factor binding sites (promoters or motifs) of genes in the Solanum Lycopersicum plant with the hope that this will give insight into those genes that are similarly regulated.

Progressive Insurance Competitive Intelligence: Analyzing Agency Data

Student Researchers: Jubilate Lema, Ye Lv, Joseph Wilch
Faculty Advisors: Pam Pierce and John Ramsay (Mathematics)

The purpose of this AMRE project was to analyze data collected from independent agents across the country to determine in what demographic groups Progressive isn’t competitive, either offering a premium too low or offering a premium too high. An analogous goal was to figure out which variables Progressive weighs differently in respect to how they set premiums compared to the competition. Our results will give Progressive Insurance a better idea of areas where they could be more successful.

Factors that Affect Student Attrition

Student Researchers: Glenn Caventer, Kazuki Kyotani, and Ryan Snyder
Faculty Advisors: R. Drew Pasteur (Mathematics) and Sofia Visa (Computer Science)

The purpose of this project was to analyze data in order to better understand student retention, specifically, students who left the College of Wooster. How can we utilize admissions data to predict the likelihood that a student will leave the College by the end of his or her first year? Using additional information attained during a student’s first semester, how well can we predict the likelihood that a student will leave the College by the end of his or her first year? Which students are most at-risk of not graduating in four years? We attempted to answer those three questions using several methods, such as probit regressions and artificial neural networks.