Event Prediction and Forecasting [top]

Event prediction is concerned with modeling where, when, and why events occur. The estimated locations and times of future events can be informed by characteristics and attributes of the spatio-temporal environment (i.e., feature space) as well as by the patterns in the event history. A Hawkes, or self-exciting point process, model is a formalized approach to combine information from both the feature space and the event history while automatically determining how much weight to give to recent events. These models have been used to represent processes such as: earthquakes, crime and conflict, traffic incidents, social media activity, financial markets, new product adoption, social network communication, neural spiking, and disease spread.

Topics include: point processes, Hawkes models, self-exciting models, kernel density estimation, spatial statistics, time series, predictive analytics, machine learning.

Recent Papers:

  • A note on the multiplicative fairness score in the NIJ recidivism forecasting challenge. Crime Science , 10(1): 1--5, 2021. (Mohler G. and Porter M.D.) [PDF]
  • Forecasting for social good. International Journal of Forecasting , : 2021. (Rostami-Tabar B., Ali M.M., Hong T., Hyndman R.J., Porter M.D., and Syntetos A.) [PDF]
  • Rotational grid, PAI-maximizing crime forecasts. Statistical Analysis and Data Mining , 11(5): 227--236, 2018. (Mohler G. and Porter M.D.) [PDF]
  • GPU accelerated MCMC for modeling terrorist activity. Computational Statistics and Data Analysis , 71: 643--651, 2014. (White G. and Porter M.D.) [PDF]
  • Modelling the effectiveness of counter-terrorism interventions. Trends and Issues in Crime and Criminal Justice , (457): 1--8, 2014. (White G., Mazerolle L., Porter M.D., and Chalk P.) [PDF]
  • Discussion of Estimating the historical and future probabilities of large terrorist events. The Annals of Applied Statistics , 7(4): 1871--1875, 2013. (Reich B.J. and Porter M.D.) [PDF]
  • Terrorism Risk, Resilience, and Volatility: A Comparison of Terrorism in Three Southeast Asian Countries. Journal of Quantitative Criminology , 29(2): 295--320, 2013. (White G., Porter M.D., and Mazerolle L.) [PDF]
  • Evaluating temporally weighted kernel density methods for predicting the next event location in a series. Annals of GIS , 18(3): 225--240, 2012. (Porter M.D. and Reich B.J.) [PDF]
  • Self-exciting hurdle models for terrorist activity. The Annals of Applied Statistics , 6(1): 106--124, 2012. (Porter M.D. and White G.) [PDF]
  • Endogenous and Exogenous Effects in Contagion and Diffusion Models of Terrorist Activity. (White G., Ruggeri F., and Porter M.D.) [PDF]
  • Contagion and Diffusion Models for the Dynamics of Terrorist Activity. [Under Contract with CRC Press] (White G.W. and Porter M.D.)

Recent Funding:

  • Ivy Foundation COVID-19 Translational Research Fund: Epidemiologic Modeling, Public Health Surveillance and Sewershed Monitoring to Predict Surges in the COVID-19 Pandemic ; $100,000.
  • Center for Advanced Public Safety: Predictive Crash Analytics ; $119,996.

Pattern and Event Detection [top]

Pattern and event detection is concerned with discovering patterns and anomalies in data. This often involves problems where the quick detection of anomalies or unusual observations from a data stream is crucial (e.g., disease and crime outbreak detection, activation of terrorist sleeper cell). But in addition to being able to detect the changes quickly, we should also be able to identify the cause of the event - like the spatial location where the outbreak started or the network neighborhood responsible for fraudulent financial activities. Clustering is another approach to making sense of complex data. This is the task of grouping observations that share features or have similar patterns. Similarly, data mining and machine learning approaches can be used to identify the features most useful for prediction or understanding the data generating process. Recent projects involve detecting anomalies in large dynamic networks and spatial event data.

Topics include: anomaly detection, dynamic network analysis, spatial hotspot detection, clustering, data mining, NMF.

Recent Papers:

  • Changing Presidential Approval: Detecting and Understanding Change Points in Interval Censored Polling Data. Stat , (): e463, 2022. (Tian J. and Porter M.D.) [PDF]
  • Learning to rank spatio-temporal event hotspots. Crime Science , 9(3): 1--12, 2020. (Mohler G., Porter M.D., Carter J., and LaFree G.) [PDF]
  • Detecting, identifying, and localizing radiological material in urban environments using scan statistics. IEEE International Symposium on Technologies for Homeland Security (HST) , 1--6, 2019. (Porter M.D. and Akakpo A.) [PDF]
  • Learning to rank spatio-temporal event hotspots. URBCOMP2018 , 2018. (Mohler G., Porter M.D., Carter J., and LaFree G.) [PDF]
  • Optimal Bayesian Clustering using Non-negative Matrix Factorization. Computational Statistics and Data Analysis , 128: 395--411, 2018. (Wang K. and Porter M.D.) [PDF]
  • How the Choice of Safety Performance Function Affects the Identification of Important Crash Prediction Variables. Accident Analysis and Prevention , 88(1): 1--8, 2016. (Wang K., Simandl J.K., Porter M.D., Graettinger A.J., and Smith R.K.) [PDF]
  • Innovative Methods for Terrorism and Counterterrorism Data. In Evidence-Based Counterterrorism Policy , Springer New York, 91--112, 2012. (Porter M.D., White G., and Mazerolle L.) [PDF]
  • Network Neighborhood Analysis. IEEE Int. Conf. on Intelligence and Security Informatics (ISI) , 31-36, 2010. (Porter M.D. and Smith R.) [PDF]
  • Mixture Likelihood Ratio Scan Statistic for Disease Surveillance. Advances in Disease Surveillance , 5: 1, 2008. (Neimi J.B., Porter M.D., and Reich B.J.) [PDF]
  • Detecting local regions of change in high-dimensional criminal or terrorist point processes. Computational Statistics \& Data Analysis , 51(5): 2753 -- 2768, 2007. (Porter M.D. and Brown D.E.) [PDF]

Recent Funding:

  • National Cooperative Highway Research Program (NCHRP): Implementing and Leveraging Machine Learning at State Departments of Transportation ; $350,000.
  • iThrive Pilot Translation and Clinical Studies and Engineering in Medicine: Impact Quantification of Donor Echocardiographic Data on Pediatric Heart Transplant Recipient Outcomes ; $82,716.
  • Centers for Disease Control and Prevention: Anomaly Detection in Space and Time ; $50,000.

Data Linkage [top]

Data linkage is concerned with matching data that correspond to the same entity. Our work in data linkage is, in essence, the assessment of statistical evidence in large disparate data. One example is crime linkage, the process of grouping together the unsolved crimes that were committed by the same offender(s). We use the behavioral patterns in crime data that have the most power to discriminate between offenders to create statistical measures of evidence. This helps crime analysts make the complex comparisons of not only the similarity, but also the distinctiveness of variables related to an offender's site selection behaviors, crime scene behaviors, and possibly even eyewitness descriptions. Besides applications in statistical forensics, concepts of data linkage can also be used for record linkage, the process of linking data records that correspond to the same entity. The recent political claims of voter fraud fit into this category as do medical record linkage to improve health outcomes and detect dangerous drug interactions. We are also developing network based approaches to road conflation, the task of merging multiple maps to better understand the traversable roads and paths of a region.

Topics include: crime linkage, statistical forensics, record linkage, road conflation, clustering, Bayesian.

Recent Papers:

  • Criminal Consistency and Distinctiveness. Systems and Information Engineering Design Symposium (SIEDS) , 1--3, 2020. (Koch A., Tian J., and Porter M.D.) [PDF]
  • Consistency and specificity in burglars who commit prolific residential burglary: Testing the core assumptions underpinning behavioural crime linkage. Legal and Criminological Psychology , 21(1): 77--94, 2016. (Bouhana N., Johnson S.D., and Porter M.D.) [PDF]
  • A Statistical Approach to Crime Linkage. The American Statistician , 70(2): 152--165, 2016. (Porter M.D.) [PDF]
  • crimelinkage: Statistical Methods for Crime Series Linkage. [R package version 0.0.4] (Porter M.D. and Reich B.J.) [PDF]
  • 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. (Reich B.J. and Porter M.D.) [PDF]

Recent Funding:

  • National Institute of Justice: Statistical Methods for Spatio-Temporal Crime Series Linkage ; $551,656.