Committee Chair

Sartipi, Mina

Committee Member

Kandah, Farah; Ward, Michael


Dept. of Computer Science and Engineering


College of Engineering and Computer Science


University of Tennessee at Chattanooga

Place of Publication

Chattanooga (Tenn.)


Vehicular accidents within Tennessee increased by 25% within 2009-2019 according to Tennessee’s Integrated Traffic Analysis Network. Accidents rank in the top three causes of accidental death across all ages in the U.S, and in 2017 accounted for 11.9% of all deaths by injury, the National Vital Statistics Report and Center for Health Statistics report. Accidents represent a massive cost in economics with 12.5 million in damages within 2018 from statistics from National Safety Council. These statistics indicate need for thorough investigation into reduction of accidents in our society. This thesis focuses on that need, with introduction of a novel predictive model based on historical accident occurrence in Hamilton County, Tennessee. The use of weather forecasts, roadway geometrics, and aggregated variables aids in creation of predictions for future accident occurrence. Additionally, an application is presented for use by local law enforcement and emergency services to assist resource deployment based upon predictions.


I would first like to thank my wife, for always being my editor, rubber ducky, and general cheerleader. I would have never gotten this far without you, and I can never thank you enough. I would also like to thank the faculty and staff at the University of Tennessee for everything that they’ve taught me, whether in the classroom or in the ’real-world’. Thank yous are also in order for my committee members, Dr. Michael Ward and Dr. Farah Kandah. An additional note of gratitude is due to Dr. Claire McCullough, who always encouraged me to do the work while also staying absolutely myself. Another thank you is in store for Grace McPherson, for introducing me to how fun and engaging computer science outside of class time could be. I would also like to thank the members of the Chattanooga City IT team, Mr. Kevin Comstock - Smart City Director of Chattanooga, the wonderful officers of the Chattanooga Police Department for their time, feedback, and input, everyone at the Enterprise Center, and NSF US Ignite for funding this project through award number 1647161. Of course, no proper listing of thank-yous would be complete without acknowledging the amazing research team I’ve spent the last three years of my life working with. Whether I met you back in the days of SCAL or more recently with the team at CUIP, I owe so much to you. A special thank you is in store for Jose Stovall and Dr. Mina Sartipi. Without my friendship with Jose, I would have never begun this amazing Data Science journey, and without Dr. Sartipi’s direction, I would certainly not be penning this thesis. I am in eternal debt to Rebekah Thompson, for her guidance on this eventful thesis adventure based on her previous journey (There and back again...). Finally, a thank you to my research partner, Jeremy Roland. He has always made sure my writing doesn’t get too loquacious, and my graphs not too detailed. Thanks Jeremy, I’ll try to keep this one brief.


M. S.; A thesis submitted to the faculty of the University of Tennessee at Chattanooga in partial fulfillment of the requirements of the degree of Master of Science.




Engineering mathematics; Traffic accidents; Traffic accident investigation


Hamilton County (Tenn.)


Machine learning; Data analysis; Neural networks; Accident prediction; Rare event analysis

Document Type

Masters theses




ix, 82 leaves