A5052235206- Advanced Statistical Process Monitoring
- Complex Network Analysis Techniques
- Scientific Measurement and Uncertainty Evaluation
- Network Security and Intrusion Detection
- Advanced Statistical Methods and Models
- Opinion Dynamics and Social Influence
- Fuzzy Systems and Optimization
- Quality and Safety in Healthcare
- Data Mining Algorithms and Applications
- Data Visualization and Analytics
- Fault Detection and Control Systems
- Information and Cyber Security
- Pesticide Residue Analysis and Safety
- Bayesian Modeling and Causal Inference
- Multi-Criteria Decision Making
- Video Analysis and Summarization
- Scientific Computing and Data Management
- Time Series Analysis and Forecasting
- Anomaly Detection Techniques and Applications
- Optimization and Mathematical Programming
- Crime Patterns and Interventions
- Advanced Graph Neural Networks
- Statistical Distribution Estimation and Applications
Virginia Tech
2014-2025
Companies routinely perform life tests on their products. Each of these typically involves testing several units simultaneously with interest in the times to failure. Two aspects often associated lifetime data that make development a control-charting procedure more demanding are tend be nonnormally distributed and censored. In this paper, one-sided lower upper likelihood-ratio—based cumulative sum (CUSUM) control charting procedures developed for Type I right-censored Weibull monitor changes...
Abstract Timely detection of anomalous events in networks, particularly social is a problem increasing interest and relevance. A variety methods have been proposed for monitoring such including the window‐based scan method by previous study. However, research assessing performance this other has sparse. In article, we use simulated network structures to study Priebe et al method. The power high only when more than half experiences behavior or if extreme. Both can be represented...
Abstract Dynamic control limits can be useful in designing charts, especially when sample sizes, risk scores, or other covariate values change over time. Computer simulation used to the conditional false alarm rate and thus in‐control run length properties. We show that this approach adaptive exponentially weighted moving average (AEWMA) charts for which chart smoothing parameter at a given time point depends on observed value point. use AEWMA as examples, but applied cumulative sum (CUSUM)...
Social networks have become ubiquitous in modern society, which makes social network monitoring a research area of significant practical importance. data consist interactions between pairs individuals that are temporally aggregated over certain interval time, and the level such temporal aggregation can substantial impact on monitoring. There been several studies effect process literature, but no We use degree corrected stochastic block model (DCSBM) to simulate anomalies analyze these...
An integral part of the design control charts, including multivariate exponentially weighted moving average (MEWMA) chart, is determination appropriate limits for prospective monitoring. Methods using Markov chain analyses, equations, and simulation have been proposed to determine MEWMA chart when are based on a specified in-control run length (ARL) value. A drawback usual approach that conditional false alarm rate (CFAR) these charts varies over time in what might be an unexpected...
The integrity of Phase II control charting depends on the accuracy I estimation. Studies have shown that extremely large sample sizes are needed in to ensure performance charts with estimated in‐control parameters is comparable known parameters. size recommendations can be impractical for attribute charts. In this article, c ‐chart an average number non‐conforming items assessed. We show sampling variability associated estimation results a high percentage run lengths well below corresponding...
Traditional statistical process monitoring (SPM) provides a useful starting point for framing and solving network problems. In this paper the panelists discuss similarities differences between two fields they describe many challenges open problems in contemporary research. The also potential outlets avenues disseminating such
In this article, the panelists broadly discuss definition of network monitoring, and how it may be similar to or different from surveillance change-point detection. The discussion uncovers ambiguity contradictions associated with these terms we argue that lack clarity is detrimental field. also describe existing emerging applications which serves illustrate wide applicability tools research
Creating an interactive, accurate, and low-latency big data visualisation is challenging due to the volume, variety, velocity of data. Visualisation options range from visualising entire dataset, which could take a long time be taxing system, small subset fast less system but also lead less-beneficial as result information loss. The main research questions investigated by this work are what effect sampling has on insight how provide guidance users in navigating trade-off. To investigate...
The study and use of network monitoring methodology is informed by its need in government, industry, technology. Here, the panelists discuss broader impacts these sectors, how development new methods influenced institutions, what challenges to be addressed next 5 10 years. There a strong consensus that sectors each play an important role innovation techniques. Applications cyber security, transportation, infectious disease monitoring, engineering, artificial intelligence are discussed.
Research in network monitoring spans a large and growing number of disciplines, including mathematics, physics, computer science, statistics. Here, the panelists discuss advantages disadvantages interdisciplinary nature area. It is largely agreed that integrating expertise from many disciplines drives innovation development, but several notable barriers are discussed limit area's full potential.