Graphical Models Research Articles

ARMA Identification of Kronecker graphical models

We address the problem to estimate a Kronecker graphical model corresponding to an autoregressive and moving average (ARMA) stochastic process. The graph describes conditional dependence relations among the components of the process. We propose a Bayesian approach to estimate the model and the topology. We test the effectiveness of the proposed method by some numerical experiments.

High-dimensional functional graphical model structure learning via neighborhood selection approach

High-dimensional functional graphical model structure learning via neighborhood selection approach

Fast Integration of Poisson Distributions for Dead Sensor Marginalization

In dual-phase time-projection chambers there are photosensor arrays arranged to allow for inference of the positions of interactions within the detector. If there is a gap in data left by a broken or saturated photosensors, the inference of the position is less precise and less accurate. As we are unable to repair or replace photosensors once the experiment has begun, we develop methods to estimate the missing signals. Our group is developing a probabilistic graphical model of the correlations between the number of photons detected by adjacent photosensors that represents the probability distribution over photons detected as a Poisson distribution. Determining the posterior probability distribution over a number of photons detected by a sensor then requires integration over a multivariate Poisson distribution, which is computationally intractable for high-dimensions. In this work, we present an approach to quickly calculate and integrate over a multidimensional Poisson distribution. Our approach uses Zarr, a Python array compression package, to manage large multi-dimensional arrays and approximates the log factorial to quickly calculate the Poisson distribution without overflow.

Accurate brain pharmacokinetic parametric imaging using the blood input function extracted from the cavernous sinus.

Brain pharmacokinetic parametric imaging based on dynamic positron emission tomography (PET) scan is valuable in the diagnosis of brain tumor and neurodegenerative diseases. For short-axis PET system, standard blood input function (BIF) of the descending aorta is not acquirable during the dynamic brain scan. BIF extracted from the intracerebral vascular is inaccurate, making the brain parametric imaging task challenging. This study introduces a novel technique tailored for brain pharmacokinetic parameter imaging in short-axis PET in which the head BIF (hBIF) is acquired from the cavernous sinus. The proposed method optimizes the hBIF within the Patlak model via data fitting, curve correction and Patlak graphical model rewriting. The proposed method was built and evaluated using dynamic PET datasets of 67 patients acquired by uEXPLORER PET/CT, among which 64 datasets were used for data fitting and model construction, and 3 were used for method testing; using cross-validation, a total of 15 patient datasets were finally used to test the model. The performance of the new method was evaluated via visual inspection, root-mean-square error (RMSE) measurements and VOI-based accuracy analysis using linear regression and Person's correlation coefficients (PCC). Compared to directly using the cavernous sinus BIF directly for parameter imaging, the new method achieves higher accuracy in parametric analysis, including the generation of Patlak plots closer to the standard plots, better visual effects and lower RMSE values in the Ki (P = 0.0012) and V (P = 0.0042) images. VOI-based analysis shows regression lines with slopes closer to 1 (P = 0.0019 for Ki ) and smaller intercepts (P = 0.0085 for V). The proposed method is capable of achieving accurate brain pharmacokinetic parametric imaging using cavernous sinus BIF with short-axis PET scan. This may facilitate the application of this imaging technology in the clinical diagnosis of brain diseases.

On the identification of ARMA graphical models

On the identification of ARMA graphical models

A Probabilistic Graphical Model for Predicting Cascade Failures of Electric Vehicle Charging Networks Caused by Hurricanes

A Probabilistic Graphical Model for Predicting Cascade Failures of Electric Vehicle Charging Networks Caused by Hurricanes

Simplifying Bayesian analysis of graphical models for the social sciences with easybgm: A user-friendly R-package

Network psychometrics models psychological constructs as interconnected variables. Rather than treating variables as independent entities, network analysis views them as nodes in a system that interact with each other; their interactions yield partial associations. Recently, researchers have empha-sized the use of Bayesian methods in graphical modeling to accurately quan-tify uncertainty in the model and its parameters. Several R packages have been developed that implement different Bayesian estimation approaches for graphical modeling in R. However, they all require different inputs and pro-duce different outputs, making them difficult to use for applied researchers. In this paper, we present a user-friendly R package called easybgm that combines the powerful analysis tools into a cohesive package for applied re-searchers. The package allows researchers to fit any type of cross-sectional data, extract results, and visualize results such as network plots, edge evi-dence plots, and structure uncertainty plots. We introduce the package and demonstrate its use with two examples.

Analyses and algorithms of personnel safety in mines using event tree and Bayesian network method

Determination of the hazard index in ensuring the safety of personnel in underground mines requires a comprehensive approach. Different methods can be applied depending on the specifics and conditions of work. This article discusses event tree and Bayesian network methods as tools for systematic analysis and assessment of personnel safety risks. The event tree method allows the sequence of events that lead to certain outcomes to be ordered and visualised. Analysing this event tree can help identify critical points and determine likely ways to prevent accidents. The Bayesian grid method is a graphical model of the probabilistic relationships between different variables. When analysing personnel safety, the Bayesian grid method can take into account the probability of events, the factors affecting safety and their interrelationships.

A Bayesian network-based framework for SHM data fusion supporting bridge management

Guidelines for Risk Classification and Management, Safety Assessment and Monitoring of Bridges, issued in 2020 by the Italian Ministry of Sustainable Mobility, promote the use of Structural Health Monitoring (SHM) systems for risk assessment of bridges. Prestressed or ordinary reinforced concrete bridges are designed to maintain their functionality over a long period of time. However, during their service life, such structures may be exposed to ever- increasing traffic volumes, extreme weather conditions and/or marine environments and so on. Therefore, constant maintenance and timely interventions are crucial to ensure the functionality of the transportation network. The current management process is mainly based on visual inspections, while the aggregation of information coming from multiple sources is still a major challenge. In this context, the main objective of the present work is to develop a general Bayesian framework capable of processing the different sources of information based on a Bayesian network (BN), a probabilistic graphical model that represents a set of variables and their conditional dependencies. A BN is a useful tool for the management of bridges since it allows the prediction of damage states and failure conditions based on the knowledge by visual inspections and SHM. The combined use of SHM and Bayesian approaches can reduce the overall risk of failure increasing the efficiency of the infrastructure system. The procedure is exemplified and validated with reference to a simply supported post-tensioned case study bridge.

Project for the development of the Eurasian transport framework in the countries of Central Asia

The paper puts the importance of an international transport corridor for the Republic of Kazakhstan. Two directions of railway communication within the latitudinal Central Eurasian Corridor have been identified. A technique has been developed for applying simulation modeling to identify the bottlenecks on the way of railway flows. Bottlenecks include the intersections of vehicular and pedestrian traffic with rail traffic. In such places, conflicting flows are formed, which may be accompanied by a merger of individual flows. Using the methods of structural analysis, synthesis, system analysis and step-by-step inclusion (or exclusion), the input parameters of simulation models of objects and processes have been established for the case study of Aktobe city in the Republic of Kazakhstan. A graphical model of the behavior of groups of objects for conflicting flows in bottlenecks has been developed, being the basis for the simulation model.

Distributed Inference With Variational Message Passing in Gaussian Graphical Models: Tradeoffs in Message Schedules and Convergence Conditions

Distributed Inference With Variational Message Passing in Gaussian Graphical Models: Tradeoffs in Message Schedules and Convergence Conditions

Resolution of cyclic dependencies of the graphical model interconnection requirements for software

Resolution of cyclic dependencies of the graphical model interconnection requirements for software

Progressive Detection of Uncertainty in Natural Language Processing Using a Labeled Variable Dimension Kalman Filter

Owing to rapid advancements in information and communication technology, natural language processing (NLP) methods have attracted considerable attention. Estimating uncertainty is a well‐established problem in NLP. Developing systems capable of automatically detecting uncertain terms within a natural language is critical for developing effective text‐analysis algorithms. Consequently, several NLP tools have been developed for this purpose. However, there are several obstacles to the design of fast and effective NLP technologies that effectively analyze ambiguous terms in natural languages. This work focuses on a particular method for detecting uncertainty, termed progressive detection, which can be accomplished using a dynamic probabilistic graphical model to solve these challenges. We iteratively upgraded the Kalman Filter to Labeled Variable Dimension Kalman Filter (LVDKF) based on observations from real datasets. The LVDKF learns two basic features from positive and negative words. The formula was then derived using a dynamic detection system, designed based on conditional probability. Finally, we simulated progressive detection experimental conditions and assessed the LVDKF on two publicly available datasets. It performs better than the baseline methods in these tests, which shows that it is effective at detecting changes over time.

Основы формирования и управления агломерациями в сельских территориях России

The article focuses on the basics of the formation and management of agglomerations in rural areas of Russia. The author’s formulation of the concept of «agglomeration in rural areas» is given. Criteria have been identified for the purpose of determining agglomerations in rural areas, the main of which include the population, the presence of a center or core of the agglomeration, the infrastructural community of municipalities in the agglomeration, the presence of agricultural production in the agglomeration, the presence of sufficiently dense buildings. Based on the criteria presented, the classification of agglomerations in rural areas is given. A model of the mechanism for managing the functioning and development of agglomerations in rural areas is presented. The main aspects of the implementation of the mechanism for managing the functioning and development of agglomerations in rural areas are substantiated, and a logical and graphical model of the phased implementation of the agglomeration management system in rural areas is presented.

Identification of the causal relationship between features and barriers of product–service systems based on Bayesian network model

Product-service systems (PSS) potentially offer a promising avenue for balancing resource efficiency improvement and economic growth to achieve sustainability. To successfully implement and operate PSS, early identification and mitigation of implementation barriers are required for enhanced social acceptability. To this end, determining the interrelationship between designing PSS features and potential barriers is crucial. Existing research focused on the PSS attributes and its barriers in isolation, falling short of a holistic understanding of their causal relationships. This paper aims to fill this gap by amalgamating case study data from literature survey and subsequently constructing a Bayesian network model. By establishing a graphical model and probability table, this study identified the stochastic dependencies between PSS features and barriers. This approach not only captures their relationship but also assists in predicting potential barriers based on PSS characteristics during the design phase. The results of this study are expected to furnish PSS designers with a tool for anticipating and addressing implementation barriers, thereby facilitating the successful implementation of PSS.

On complexity of modeling the business process lifecycle

Business Process LifeCycle (BPLC) is a topic never disappearing from the discussions in the Business Process Management (BPM) research. Nevertheless, most graphical models of BPLC are simple and need extended textual explanations to clarify them. What is more, to the best of our knowledge, these models are separated from the target process that is being changed in the BPLC. The paper investigates a possibility to have a model where lifecycle activities are connected to the elements of the target process, and assess the practical usefulness of such a model. The research uses the Fractal Enterprise Modeling (FEM) technique for building the model. The paper presents a generic model of BPLC and its extension for one step, and discusses potential usefulness of the results achieved so far.

Simulation modeling as a means of integrating infrastructure for individual mobility in urban environments

The article presents the concept of sustainable development of the territory of the Moscow city district by introducing paths for individual mobility into the transport network. A simulation model of the street road network is created to substantiate the efficiency of implementation, traffic safety, and connectivity of street routes. The accuracy and validity of the obtained model is determined by the use of research methods, which provide a detailed description of the problem area, structured processes and objects, and also establish features of their interaction. To visualize the obtained results, a graph of conflicting and connecting flows in the transport network of the district and graphical models of their behavior are developed.

Dietary Networks Identified by Gaussian Graphical Model and Odds of Major Depressive Disorder: A Case‐Control Study

Dietary Networks Identified by Gaussian Graphical Model and Odds of Major Depressive Disorder: A Case‐Control Study

Learning conditional dependence graph for concepts via matrix normal graphical model

Learning conditional dependence graph for concepts via matrix normal graphical model

A glimpse into the future: revealing the key factors for survival in cognitively impaired patients.

Drawing on prospective data from the National Health and Nutrition Examination Survey (NHANES), our goal was to construct and validate a 5-year survival prediction model for individuals with cognitive impairment (CI). This study entailed a prospective cohort design utilizing information from the 2011-2014 NHANES dataset, encompassing individuals aged 40 years or older, with updated mortality status as of December 31, 2019. Predictive models within the derivation and validation cohorts were assessed using logistic proportional risk regression, column-line plots, and least absolute shrinkage and selection operator (LASSO) binomial regression models. The study enrolled a total of 1,439 participants (677 men, mean age 69.75 ± 6.71 years), with the derivation and validation cohorts consisting of 1,007 (538 men) and 432 (239 men) individuals, respectively. The 5-year mortality rate stood at 16.12% (n = 232). We devised a 5-item column-line graphical model incorporating age, race, stroke, cardiovascular disease (CVD), and blood urea nitrogen (BUN). The model exhibited an area under the curve (AUC) of 0.772 with satisfactory calibration. Internal validation demonstrated that the column-line graph model displayed strong discrimination, yielding an AUC of 0.733, and exhibited good calibration. To sum up, our study successfully developed and internally validated a 5-item nomogram integrating age, race, stroke, cardiovascular disease, and blood urea nitrogen. This nomogram exhibited robust predictive performance for 5-year mortality in individuals with CI, offering a valuable tool for prognostic evaluation and personalized care planning.