About

J. Eric Dietz, PhD, PE, is a Professor and Interim Department Head at Purdue University's College of Technology. His research interests include optimization of emergency response, homeland security and defense, energy security, and engaging veterans in higher education. As a Director in Purdue’s Discovery Park, he is responsible for catalyzing homeland security research, increasing the societal impact of Purdue research, and organizing interdisciplinary projects within the university. Prior to his current role, Dr. Dietz served as the founding Executive Director for the Indiana Department of Homeland Security, where he was responsible for emergency planning, training, fire and building safety, and disaster response for Indiana. During this period, he led Indiana’s response to multiple Presidential Major Disasters and Emergency Declarations, and was instrumental in creating the Indiana Intelligence Fusion Center and the Indiana Fire Training System. He retired as a Lieutenant Colonel from the U.S. Army in 2004, leading various Army acquisition and research programs related to power systems, chemical sensors, and command and control systems. Dr. Dietz holds a Bachelor of Science and a Master of Science in Chemical Engineering from Rose-Hulman Institute of Technology, and a PhD in Chemical Engineering from Purdue University. His professional achievements include awards such as the Sagamore of the Wabash, the President’s Award from the Indiana Fire Chiefs Association, and recognition as an Honor Alumnus of Rose-Hulman. He is actively involved in industry engagement, research, and professional affiliations, including serving as a Director at the Purdue Homeland Security Institute.

Research topics

• Computer Science
• Computer Security
• Data Mining
• Artificial Intelligence
• Chemistry
• Operating system
• Real-time computing
• Telecommunications
• Chemical engineering
• Systems engineering
• Embedded system
• Materials science
• Engineering
• Database
• Data science
• Composite material

Selected publications

• P-1560. Risk Factors for Extended-Spectrum β-Lactamase-Producing <i>Escherichia coli</i> Urinary Tract Infections

  Open Forum Infectious Diseases · 2025-01-29

  articleOpen access

  Abstract Background E. coli is the leading cause of urinary tract infections (UTIs) worldwide. Antibiotic resistance among E. coli is increasing. We performed a case-control study to identify risk factors predictive of a positive extended-spectrum beta-lactamase (ESBL)-producing E. coli (ESBL-E) compared to a non-ESBL-E urine culture. Table 1 Demographic and clinical characteristics of patients with positive E. coli urine culture, Monroe County, August 2023 Methods Positive E. coli urine cultures were identified via laboratory and population-based surveillance in Monroe County as part of the CDC Emerging Infections Program in August 2023. Cases were defined as a county resident with a positive ESBL-E urine culture (resistant to ≥1 3rd-generation cephalosporin and nonresistant to carbapenems). Controls had a non-ESBL-E urine culture. All ESBL-E cases and a random sample of non-ESBL-E controls underwent medical record review to collect information on prior antimicrobial use, recent hospitalization, prior ESBL-E infection, underlying conditions, and prior long term care stay. Charlson Comorbidity Index (CCI) was calculated and categorized as low (0), middle (1-2), and high (3+). Multivariable logistic regression was used to identify independent risk factors. Results We identified 2142 patients with an E. coli urine culture; 454 underwent chart review (64 ESBL-E, 390 non-ESBL-E). Patient characteristics are summarized in Table 1. Overall, patients were mostly female (88.3%) with a median age of 61 years (IQR 39-76). Cases more commonly received antibiotics in the 30 days prior to culture (34.4% vs. 19.7%, p = 0.007), had a hospitalization in the 90 days prior to culture (17.2% vs. 8.5%, p = 0.029), had prior ESBL-E culture (25.0% vs. 3.3%, p &amp;lt; 0.0001), a high CCI (32.8% vs. 19.7%, p = 0.006), prior LTC stay (18.8% vs. 6.4%, p = 0.0008), and diabetes (35.9% vs. 19.5%, p=0.003). After adjusting for confounding factors, only prior ESBL-E culture (OR = 8.22; 95% CI: 3.68-18.40) and diabetes (OR = 2.32; 95% CI: 1.31-4.09) were significantly predictive of the outcome. Conclusion Patients with ESBL-E are more medically complex with increased healthcare exposure compared to patients with non-ESBL-E. Presence of a prior ESBL-E culture and diabetes were specific risk factors. Future studies are needed to better understand the collective impact of ESBL-E drivers to inform prevention strategies. Disclosures All Authors: No reported disclosures

  PublisherOA PDFDOI

• FraudX SimS: A Synthetic Dataset for Anomaly Detection in Payment-Card Transactions

  IEEE Access · 2025-01-01

  articleOpen access

  Progress in detecting payment fraud is challenged by the limited variety of publicly available datasets. Relying on one or two datasets makes it hard to compare fairly, disguises sensitivity to data changes, and limits the ability to evaluate explainable methods in depth. This article introduces FraudX SimS, a scenario-labeled synthetic dataset designed to expand the set of benchmarks for anomaly detection in payment transactions, particularly in the context of fraud detection. The dataset preserves the class imbalance between legitimate and fraudulent activity and includes openly specified spatial, temporal, and behavioral features, allowing direct application of explainable artificial intelligence techniques. We establish baselines with standard machine learning models and report accuracy, precision, recall, F1-score, confusion-matrix results, and the area under the receiver operating characteristic curve (AUC-ROC) and the area under the precision–recall curve (AUC-PR), with a primary emphasis on recall given the cost of missed fraud. We further employ Shapley additive explanations to quantify feature contributions, enabling transparent error analysis and model refinement. Although synthetic, the dataset is constructed to support reproducible experimentation and cross-study comparisons, advancing the development of reliable and interpretable fraud-detection methods.

  PublisherDOI

• P-292. Epidemiology of <i>Escherichia coli</i> from Urine Sources at Three U.S. Sites, August 2023

  Open Forum Infectious Diseases · 2025-01-29

  articleOpen access

  Abstract Background Escherichia coli is the most frequent cause of urinary tract infections (UTIs) in the US, but surveillance for urinary E. coli (uEC) is often limited to multidrug-resistant strains. The CDC’s Emerging Infections Program piloted one month of active population- and laboratory-based surveillance for uEC of any phenotype in 3 U.S. sites (representing &amp;gt;3.4 million people) to better characterize its epidemiology and burden. Methods We defined an incident uEC case as the first E. coli isolated from urine in a resident of the surveillance area during August 2023. Demographic data were obtained for all incident cases. Approximately 400 cases per site were randomly selected for chart review. One-month incidence rates were generated using 2022 US census data for participating surveillance areas. Cases were classified as community-associated (CA; no healthcare risk factors); hospital-onset (HO; urine obtained ≥ 3 days after hospital admission), or healthcare-associated, community onset (HACO; urine obtained in a non-hospital setting or &amp;lt; 3 days after hospital admission with healthcare exposures in the prior year). Results Of 5241 uEC patients, 86.1% were female; median age was 60 years (IQR 37–76). The 1-month incidence rate was 151.7 per 100,000 population and greater among those aged ≥ 60 vs &amp;lt; 60 years (362.7 vs 94.0). Of 1197 patients with data from chart review, 1104 (92.2%) had cultures collected as an outpatient; 732 (61.2%) had an underlying comorbidity reported; 1113 (93.0%) had a documented UTI diagnosis; 269 (22.5%) reported recurrent UTI; and 43 (3.6%) developed E. coli bacteremia. No signs or symptoms associated with uEC were reported for 372 (31.1%) patients. In total, 287 (24%) were hospitalized within 30 days after urine culture collection; 51/287 (17.8%) had a urinary catheter in place 2 days before culture collection. Most uEC were CA (867; 72.4%), followed by HACO (325; 27.2%) and HO (3; 0.3%). Of 3289 uEC isolates with local testing results, 255 (7.8%) were extended-spectrum β-lactamase producing and 5 (&amp;lt; 1%) were carbapenem-resistant. Conclusion These findings emphasize the high burden of uEC in these communities, notably among older adults and females. Effective prevention strategies are needed to decrease uEC-associated morbidity. Disclosures All Authors: No reported disclosures

  PublisherDOI

• A Systematic Review of Machine Learning in Credit Card Fraud Detection Under Original Class Imbalance

  Computers · 2025-10-15 · 12 citations

  reviewOpen access

  Credit card fraud remains a significant concern for financial institutions due to its low prevalence, evolving tactics, and the operational demand for timely, accurate detection. Machine learning (ML) has emerged as a core approach, capable of processing large-scale transactional data and adapting to new fraud patterns. However, much of the literature modifies the natural class distribution through resampling, potentially inflating reported performance and limiting real-world applicability. This systematic literature review examines only studies that preserve the original class imbalance during both training and evaluation. Following PRISMA 2020 guidelines, strict inclusion and exclusion criteria were applied to ensure methodological rigor and relevance. Four research questions guided the analysis, focusing on dataset usage, ML algorithm adoption, evaluation metric selection, and the integration of explainable artificial intelligence (XAI). The synthesis reveals dominant reliance on a small set of benchmark datasets, a preference for tree-based ensemble methods, limited use of AUC-PR despite its suitability for skewed data, and rare implementation of operational explainability, most notably through SHAP. The findings highlight the need for semantics-preserving benchmarks, cost-aware evaluation frameworks, and analyst-oriented interpretability tools, offering a research agenda to improve reproducibility and enable effective, transparent fraud detection under real-world imbalance conditions.

  PublisherDOI

• FraudX AI: An Interpretable Machine Learning Framework for Credit Card Fraud Detection on Imbalanced Datasets

  Computers · 2025-03-25 · 22 citations

  articleOpen access

  Credit card fraud detection is a critical research area due to the significant financial losses and security risks associated with fraudulent activities. This study presents FraudX AI, an ensemble-based framework addressing the challenges in fraud detection, including imbalanced datasets, interpretability, and scalability. FraudX AI combines random forest and XGBoost as baseline models, integrating their results by averaging probabilities and optimizing thresholds to improve detection performance. The framework was evaluated on the European credit card dataset, maintaining its natural imbalance to reflect real-world conditions. FraudX AI achieved a recall value of 95% and an AUC-PR of 97%, effectively detecting rare fraudulent transactions and minimizing false positives. SHAP (Shapley additive explanations) was applied to interpret model predictions, providing insights into the importance of features in driving decisions. This interpretability enhances usability by offering helpful information to domain experts. Comparative evaluations of eight baseline models, including logistic regression and gradient boosting, as well as existing studies, showed that FraudX AI consistently outperformed these approaches on key metrics. By addressing technical and practical challenges, FraudX AI advances fraud detection systems with its robust performance on imbalanced datasets and its focus on interpretability, offering a scalable and trusted solution for real-world financial applications.

  PublisherOA PDFDOI

• Inaugural Defense and Security Research Symposium of the Purdue Military Research Institute

  Purdue University Press eBooks · 2024-01-01

  bookOpen access

  This document is the full conference proceedings from June 26-27, 2023.

  PublisherOA PDFDOI

• Software library for agent-based modeling and simulation of active shooter events

  Journal of Emergency Management · 2022-03-01

  articleSenior author

  Software libraries have been used for decades to produce code in a quick and cost-effective manner. The use of well-designed libraries permits software developers and other professionals to create applications due in part to code reusability. Also, good libraries grant lesser skilled developers the opportunity to make high-quality applications they otherwise could not produce. In the field of active shooting incident (ASI) research, various tools have been used for years that give researchers the ability to conduct exploratory research. However, as good as these tools might be, there has been little thought about reusability of these models and associated code. This has hindered the proper advancement of the research field given that researchers must often start from nothing when building a new model. Constant repetition of the same basic tasks has not enabled researchers to expand model fidelity and has limited time to dedicate toward the problem set. This paper proposes the creation of a new agent-based ASI library, made for the AnyLogic® system. The library assists researchers in quickly creating models using a well-defined application programming interface. It also abstracts away implementation details so that the library user does not get waylaid in development. The authors also recreated parts of a large and powerful AnyLogic model to observe the resultant library employment. It was determined that a significant amount of time can be saved building new models, even with the initial version of the library implementation.

  PublisherDOI

• Active shooter mitigation strategies in small rural churches

  Journal of Emergency Management · 2022-03-01 · 2 citations

  articleSenior author

  Gun violence in places of worship (POW) has long been an issue and has been addressed repeatedly in the literature. Contextually, most of the research has been pertinent to relatively large POW, situated in an urban setting. However, rural churches have not been addressed, and they appear to have a far less defensive posture, mainly because of their remote location and the extended time required for first responders to arrive, which in turn requires a higher level of independent operation in terms of defense and medical response. Having retained an off-duty officer is a strong deterrent and provides the ability to handle any issues that may arise, including lower violence level events. If retaining an officer is not an option, having a well-trained volunteer armed team and a clear plan of action is vital to surviving such an event. Furthermore, due to the extended response and transport time, it is critical to have proper medical training, such as Stop the Bleed® and cardio-pulmonary resuscitation. This article's focus is not only on response but preparedness, which reinforces the response, as well as prevention and deterrence. An exhaustive best practices review has informed the solutions offered, supplemented by experience and recommendations of a highly experienced physical security expert and a police officer member of a Special Weapons and Tactics (SWAT) team.

  PublisherDOI

• Optimizing Cybersecurity Budgets with AttackSimulation

  2022-11-14 · 5 citations

  articleSenior author

  Modern organizations need effective ways to assess cybersecurity risk. Successful cyber attacks can result in data breaches, which may inflict significant loss of money, time, and public trust. Small businesses and non-profit organizations have limited resources to invest in cybersecurity controls and often do not have the in-house expertise to assess their risk. Cyber threat actors also vary in sophistication, motivation, and effectiveness. This paper builds on the previous work of Lerums et al., who presented an AnyLogic model for simulating aspects of a cyber attack and the efficacy of controls in a generic enterprise network. This paper argues that their model is an effective quantitative means of measuring the probability of success of a threat actor and implements two primary changes to increase the model's accuracy. First, the authors modified the model's inputs, allowing users to select threat actors based on the organization's specific threat model. Threat actor effectiveness is evaluated based on publicly available breach data (in addition to security control efficacy), resulting in further refined attack success probabilities. Second, all three elements - threat effectiveness, control efficacy, and model variance - are computed and evaluated at each node to increase the estimation fidelity in place of pooled variance calculations. Visualization graphs, multiple simulation runs (up to 1 million), attack path customization, and code efficiency changes are also implemented. The result is a simulation tool that provides valuable insight to decision-makers and practitioners about where to most efficiently invest resources in their computing environment to increase cybersecurity posture. AttackSimulation and its source code are freely available on GitHub.

  PublisherDOI

• Agent-based modeling for theme park evacuation

  Journal of Emergency Management · 2022-03-01 · 2 citations

  articleSenior author

  Each year theme parks can see up to 20 million patrons, but often little effort is put into planning for an emergency evacuation. In this study, we built a multiagent simulation model using AnyLogic® 8.5.1. The model was based on a preliminary design of a theme park provided by AOA Builds, Orlando. This research had two goals: the first was to compare evacuation time when the park is full (1) using only the main guest gate and (2) using all seven available exits. The second goal was to model first responder response time between various start and end locations within the park. Using only the main gate, evacuation took an average of 14 minutes and 51 seconds. Using all seven gates results in an average evacuation time of 11 minutes and 58 seconds. This was due to a gate being overwhelmed causing a delay in overall evacuation time. If that gate is not included in the calculation, the average evacuation time drops to 6 minutes and 44 seconds. For the purpose of measuring response times, four starting locations were chosen with the guidance of a subject matter expert. These locations included response teams positioned at the front gate, at a police station, at the service area behind a main attraction, and mobile patrol walking around the park. Based on our testing, walking around the park was the best option in terms of response time, using the main gate was 53.7 percent faster than other options and, using all seven gates, was 60.7 percent faster during an evacuation using all seven exits.

  PublisherDOI

Frequent coauthors

• Adam Kirby

  17 shared

• Eric T. Matson

  Purdue University System

  17 shared

• Joseph F. Pekny

  Purdue University West Lafayette

  12 shared

• Clifford Wojtalewicz

  United States Department of Homeland Security

  11 shared

• Byung‐Cheol Min

  Purdue University West Lafayette

  10 shared

• John Springer

  Purdue University West Lafayette

  9 shared

• Eric T. Matson

  9 shared

• Vamshi Krishna Bolly

  Purdue University System

  9 shared

Awards & honors

• Sagamore of the Wabash Award - Highest Award for Distinguish…
• President's Award - Indiana Fire Chief's Association (2006)
• Honor Alumnus - Rose-Hulman Institute of Technology