Research

University of Texas at Tyler

Home  |  Research  |  Publications  | People

 

The Data Analytics Lab is a purpose-built facility capable of performing sophisticated analysis on big data using statistics, machine learning, visualization and other tools.  Our current projects include:

 

Health Informatics

The TylerADE Project uses machine learning on the FDA's Adverse Effects Reporting System to improve prescription drug safety by identifying previously unknown but potentially dangerous drug combinations.  By analyzing the combinations of several thousand drugs across millions of historical records, the TylerADE System will be able to compute a predicted drug severity score that can be used by pharmacists and physicians at the point of prescription to determine potential health hazards.

 

Textual/Financial Prediction

Textual/Financial Prediction is the melding of textual financial news articles and machine learning to create a new class of high-frequency stock trading engine (HFT) that can recognize and adapt to immediate market changes.  Knowing that major traders are now using HFT systems that rely on textual information, our newest area is to evaluate whether we can manipulate stock prices by adjusting press releases.  Our system looks at three aspects; the words used in a news release, market timing and media outlet for release, to build a type of Press Release Engineering system that companies can use to manage their stock price.

 

Sports Analytics

Sports Analytics is the mining of relevant data from Sports-related databases to produce forecasts.  Projects include predicting winning in the NFL using a sentiment analysis of Twitter and technical analysis strategies borrowed from Wall Street, to determine relative public sentiment trajectories as a simple time-series signal.  This project has shown unexpected promise that further technical trading techniques might also be ported into the sports wagering markets to improve odds-setting as well as more accurate predictions.