51ÔÚÏß

Muma College of Business Professors, Students Take Top Honors at WITS 2016 Conference

By Keith Morelli

TAMPA (Dec. 21, 2016) -- A pair of 51ÔÚÏß Muma College of Business professors won prestigious awards at the annual Workshop on Information Technology and Systems conference, a gathering of information systems faculties from around the world. The annual conference offers an opportunity for researchers and scientists to share their latest research in the field of wireless technologies.

Kaushik Dutta and Balaji Padmanabhan, both of whom are professors and researchers in data analytics in the Information Systems Decision Sciences Department, came home with top honors from the conference held in Dublin, Ireland.

Dutta, along with PhD candidate Shalinda Adikari and associate professor Jungpil Hahn, both at the National University of Singapore, penned the paper, "Temporal Feature Grouping Based Campaign Optimization in Real-Time Bidding Digital Advertising," which won the best paper at the conference.

The paper deals with mobile marketing, which is one of the fastest growing digital marketing areas today. In mobile marketing 90 percent of the advertisement placement happens in real-time through a system called real-time bidding exchange, Dutta said.

"In this real-time bidding ecosystem, advertisers try to have maximum return of their ad investment through maximum clicks with minimum impressions," he said. "We have devised an algorithmic strategy that can automate this process (campaign optimization) for advertisers, which is still done manually. We have demonstrated the applicability of this approach in working with a major mobile advertisement tech company."

Dutta has published several other papers in the area of digital advertisement technologies in last few years.

Dutta is an associate professor with 19 years of professional and research experience in the field of enterprise IT infrastructure, data analytics and big data systems.

Padmanabhan and three others, including Muma College of Business graduate student Zachary Kazanski, won the best prototype award at the conference. Two other collaborators, Varol Kayhan and Alison Watkins – the lead writers of the prototype – are professors of Information Systems and Decision Sciences at USF St. Petersburg.

Their project was "Predicting the Winner of an NBA Game in Real-Time: A Data Snapshot Approach."

"Many systems exist today for predicting the outcomes of sports games prior to the start of the game," Padmanabhan said. "But what about during the game, in real time?"

He and Kazanksi teamed with Kayhan and Watkins from the Kate Tiedemann College of Business at USF's St. Petersburg campus, to build a system for real-time predictions of the outcomes of NBA basketball games. Their system uses a "data snapshots" approach to compare real-time game scenarios with relevant prior games going back 10 years to construct real-time predictions with game-winning probabilities that can be communicated through Twitter.

Where does real-time game data originate?

The authors wrote a text extraction code that pulls information from real-time, play-by-play data on popular sports sites, Padmanabhan said. Using historical data from over 10 years, the prototype showed good prediction accuracies when applied to a single year's games.

The researchers plan to work on improving the model in 2017 and launch a real-time prediction on Twitter sometime during the 2017-2018 NBA season.

Padmanabhan, who has worked in the analytics arena for 20 years, is director of the Center for Analytics and Creativity and Anderson Professor of Global Management at USF's Muma College of Business.