University of Tennessee

Melton Scholars Update

September 20, 2019

The Melton Scholars program pairs accomplished undergraduate students with faculty members in the Department of Business Analytics and Statistics to pursue academic research. Projects undertaken by the first and second cohorts of the program include predicting the outcome of sporting events, improving retail delivery systems, analysis of Google’s AdWords algorithm and more.

Second Cohort

Brandon Bell (Statistics) Under Michel Ballings, Bell used neural networks (an analytics algorithm) to extract audio and video of sales calls to predict whether the call would end in a sale. He also applied reinforcement learning to “teach” a neural network how to play video games such as Pong, beating human players in the process. After Bell’s graduation, Ballings hired him as a research assistant. Bell likely will join the department’s PhD program when new students are admitted.

Brian Haney (Statistics) One of Haney’s passions is sports analytics. Under the guidance of Haileab Hilafu, he studied basketball statistics to develop a model, based on live game time and team data, which gives real-time predictions of the probability that a team will win a game. 

Shelby Keller (Business Analytics, with Marketing collateral) Keller pursued a project in healthcare analytics under Bogdan Bichescu. She studied the controversial new “two midnight rule” for classifying whether a patient is inpatient or outpatient, a critical distinction in billing. Studying the records of more than 125,000 patients, she found that implementation of the rule decreased the number of inpatients admitted through the emergency department, but increased the average length of stay and charges per day. 

Austin Kuhr (Business Analytics and Supply Chain Management) Kuhr won a 2019 EURēCA prize (Exhibition of Undergraduate Research and Creative Achievement), in which each undergraduate college selects and awards the prize to their top students. He worked with Chung Eun Lee, using quantile regression and time series models, to help Colgate-Palmolive predict replenishment volumes in its supply chain. His EURēCA presentation applied these techniques to predict house prices.

Scott Lawson (Business Analytics, with Information Management collateral) Lawson collaborated with Sean Willems on data from a grocery distributor, to help manage the distributor supply chain and to best budget driver hours. They developed a means to predict how long a driver will take to deliver goods at a particular stop, given the inventory’s weight, number of items, totes and dollar amount. Lawson is now in the master’s dual degree program, earning both an MSBA and an MBA.

Marc Muszik (Accounting, with Business Analytics collateral) Muszik has a keen interest in the application of statistical and machine learning methods to financial markets. Working with Adam Petrie, one of his lines of inquiry is exploring how stock prices change before an earnings announcement. His goal is to discover signals that allow for advantageous short-term trading.

Harris Taylor (Business Analytics, with Marketing collateral) Taylor worked with Adam Petrie and Reviewbox, a Knoxville-based analytics company that helps businesses sell products on Amazon. He used data “scraped” from Amazon’s website to determine what is required to win (and keep) a product’s “buy box” (the seller associated with the “Add to Cart” and “Buy Now” buttons on a product page). 

First Cohort 

Nick Brown worked with Michel Ballings on developing a graphical interface that allows Foursquare users to find places to eat, drink and shop. He now works as a data scientist at Pilot Flying J.

Isaac Edmiston worked with Bill Seaver on a project using multivariate methods to assess student performance and detect cheating in statistics classes. He is currently a student in the MSBA program. 

Elizabeth Nichols is currently a student in the MSBA program. In her project, she collaborated with Wenjun Zhao to explore ways to generate artificial transactional data that mimics the real thing. Since most businesses do not want to share such data, and yet realistic data is required to improve algorithms, the research is valuable to the field.  

Nainika Sudheendra worked with Sean Willems on parts forecasting (patterns and demand) for Intel. She is now in Germany pursuing a graduate degree. 

Thao Tran worked with Julie Ferrara and Michel Ballings to analyze the algorithm Google Adwords uses to devise quality scores (the metric used to determine ad pricings). She is currently a student in the MSBA program.

Joe Trice worked with Bogdan Bichescu on healthcare data and studied workload allocation among doctors. He is currently a student in Haslam’s Master of Accountancy program. 

About the Program

Established in 2017, the Melton Scholars program is made possible through a generous contribution by George Melton (HCB, ’68). Melton, a Tennessee native born in Maryville, is a graduate of Red Bank High School in Chattanooga. 

After serving in the U.S. Army Signal Corp, including a year in Vietnam, Melton enjoyed a successful 43-year career in the aerospace and defense industry, which included executive positions with Allied-Signal, Tracor, BAE Systems, EG&G and Wyle, Inc.

By providing the opportunity for undergraduates to form meaningful connections with faculty members, Melton supports the university and enriches students’ lives.