Bio

B.S., Purdue University, 1998M.S., Vanderbilt University, 2003Ph.D., University of Virginia, 2006

"Never trust to general impressions, my boy, but concentrate yourself upon details."

Sherlock Holmes, in A Case of Identity

Michael D. Porter is an Associate Professor of Systems Engineering with a joint appointment at the Darden School of Business at the University of Virginia. 

Dr. Porter's research focuses on event prediction, pattern and anomaly detection, and data linkage. He has developed self-exciting point process models for processes where the occurrence of certain events can trigger a cascade of subsequent events. These models have been applied to crime, terrorism, social media, and crash data and led to a winning performance in NIJ's Real-time Crime Forecasting Challenge. Porter has also developed methodology and software for crime linkage, a type of data linkage problem where the goal is to group together unsolved crimes that were committed by the same offender(s) using the behavioral patterns obtained from crime data. 

Dr. Porter teaches courses in data science, statistics, and analytics with a particular emphasis on data-driven decision-making and reproducible research. His team has experience working with data from point processes, dynamic networks, spatial processes and GIS, time series, sensors, and social media; using methodology from data science, statistics, stochastic processes, machine learning and data mining, and analytics; to solve problems related to criminology, transportation, terrorism, defense, security, forensics, business, marketing, social media, and finance.

Prior to joining UVA in 2018, Dr. Porter was an associate professor of Applied Statistics at the University of Alabama, post-doc at North Carolina State University and SAMSI, and has industry experience as principal research scientist at DigitalGlobe/GeoEye/Spadac and project engineer at Sanford/Newell Brands.

Research Interests

  • Event Prediction and Forecasting: point processes, Hawkes models, self-exciting models, kernel density estimation, spatial statistics, time series, predictive analytics, machine learning.
  • Pattern and Event Detection: anomaly and hotspot detection, dynamic network analysis, clustering, data mining, NMF.
  • Data Linkage: crime linkage, statistical forensics, record linkage, road conflation, clustering, Bayesian modeling.
  • Machine Learning
  • Computational Statistics and Statistical Modeling
  • Data Science

Selected Publications

  • Rotational grid, PAI-maximizing crime forecasts. Statistical Analysis and Data Mining ABS Mohler G. and Porter M.D.
  • Optimal Bayesian Clustering using Non-negative Matrix Factorization. Computational Statistics and Data Analysis, 128: 395–411, 2018. ABS Wang K. and Porter M.D.
  • Learning to rank spatio-temporal event hotspots. URBCOMP2018, 2018. ABS Mohler G., Porter M.D., Carter J., and LaFree G.
  • A Statistical Approach to Crime Linkage. The American Statistician, 70(2): 152–165, 2016. ABS Porter M.D.
  • How the Choice of Safety Performance Function Affects the Identification of Important Crash Prediction Variables. Accident Analysis & Prevention, 88(1): 1–8, 2016. ABS Wang K., Simandl J.K., Porter M.D., Graettinger A.J., and Smith R.K.
  • Partially-supervised spatiotemporal clustering for burglary crime series identification. Journal of the Royal Statistical Society: Series A (Statistics in Society), 178(2): 465–780, 2015. ABS Reich B.J. and Porter M.D.
  • GPU accelerated MCMC for modeling terrorist activity. Computational Statistics & Data Analysis, 71: 643–651, 2014. ABS White G. and Porter M.D.
  • Evaluating temporally weighted kernel density methods for predicting the next event location in a series. Annals of GIS, 18(3): 225–240, 2012. ABS Porter M.D. and Reich B.J.
  • Self-exciting hurdle models for terrorist activity. The Annals of Applied Statistics, 6(1): 106–124, 2012. ABS Porter M.D. and White G.
  • Network Neighborhood Analysis. IEEE Int. Conf. on Intelligence and Security Informatics (ISI), 31-36, 2010. ABS Porter M.D. and Smith R.
  • Mixture Likelihood Ratio Scan Statistic for Disease Surveillance. Advances in Disease Surveillance, 5: 1, 2008. ABS Neimi J.B., Porter M.D., and Reich B.J.
  • Detecting local regions of change in high-dimensional criminal or terrorist point processes. Computational Statistics & Data Analysis, 51(5): 2753 – 2768, 2007. ABS Porter M.D. and Brown D.E.
Google Scholar
ResearchGate

Courses Taught

  • SYS 6018 - Data Mining
  • GBUS 7600 - Data Analysis and Optimization