The Paper Feed

Bayesian networks have been widely used for classification problems. These models, structure of the network and/or its parameters (probability distributions), are usually built from a data set. Sometimes we do not have information about all the possible values of the class variable, e.g. data about a reactor failure in a nuclear power station. This problem is usually focused as an anomaly detection problem. Based on this idea, we have designed a decision support system tool of general purpose.

The Ecosystem Management Decision Support (EMDS) system has been further enhanced with an analytical engine—BayesFusion’s SMILE (Structural Modeling, Inference, and Learning Engine) that comes with the GeNIe (Graphical Network Interface) software—for creating Bayesian network (BN) models. BNs are graphical networks of variables linked by probabilities, that have proven useful in decision-aiding for risk analysis and risk management.

Decreasing sedentary behaviour (SB) has emerged as a public health priority since prolonged sitting increases the risk of non-communicable diseases. Mostly, the independent association of factors with SB has been investigated, although lifestyle behaviours are conditioned by interdependent factors. Within the DEDIPAC Knowledge Hub, a system of sedentary behaviours (SOS)-framework was created to take interdependency among multiple factors into account. The SOS framework is based on a system approach and was developed by combining evidence synthesis and expert consensus. The present study conducted a Bayesian network analysis to investigate and map the interdependencies between factors associated with SB through the life-course from large scale empirical data.

The hierarchy of propositions has been accepted amongst the forensic science community for some time. It is also accepted that the higher up the hierarchy the propositions are, against which the scientist are competent to evaluate their results, the more directly useful the testimony will be to the court. Because each case represents a unique set of circumstances and findings, it is difficult to come up with a standard structure for evaluation. One common tool that assists in this task is Bayesian networks (BNs). There is much diversity in the way that BN can be constructed. In this work, we develop a template for BN construction that allows sufficient flexibility to address most cases, but enough commonality and structure that the flow of information in the BN is readily recognised at a glance. We provide seven steps that can be used to construct BNs within this structure and demonstrate how they can be applied, using a case example.

In court, it is typical for biological evidence to be reported at a level that only addresses how likely the DNA evidence is if it originated from a particular individual, or individuals. However, there are other questions that could be considered that would be of value in enabling the court, including the jury, to make better informed decisions. For example, although answers to specific questions such as: “Which type of bodily fluid has the DNA originated from?” or, “How was the DNA deposited at the scene?” would be probabilistic in nature, they can be crucial to the outcome of a case. The relationship between the DNA evidence, the source of the DNA and the activity that took place is described in a term called the “hierarchy of propositions.” Currently, such questions are usually answered by scientists subjectively with little to no logical framework to assist them. Bayesian networks have proven to be beneficial in providing logical reasoning by way of a likelihood ratio to help combine subjective, yet, experience‐based, opinions of experts with experimental data when answering questions which can be both complex and uncertain. These networks offer a framework that provides balance, transparency, and robustness in the evaluation of evidence. A current limitation of the use of Bayesian networks includes a lack of understanding of the underlying concepts from both forensic scientists and the courts and consequently a reduced recognition of the potential strengths.

Modelling cultural ecosystem services is challenging as they often involve subjective and intangible concepts. As a consequence they have been neglected in ecosystem service studies, something that needs remedying if environmental decision making is to be truly holistic. We suggest Bayesian Networks (BNs) have a number of qualities that may make them well-suited for dealing with cultural services. For example, they define relationships between variables probabilistically, enabling conceptual and physical variables to be linked, and therefore the numerical representation of stakeholder opinions. We assess whether BNs are a good method for modelling cultural services by building one collaboratively with canoeists to predict how the subjective concepts of fun and danger are impacted on by weir modification. The BN successfully captured the relationships between the variables, with model output being broadly consistent with verbal descriptions by the canoeists. There were however a number of discrepancies indicating imperfect knowledge capture. This is likely due to the structure of the network and the abstract and laborious nature of the probability elicitation stage. New techniques should be developed to increase the intuitiveness and efficiency of probability elicitation. The limitations we identified with BNs are avoided if their structure can be kept simple, and it is in such circumstances that BNs can offer a good method for modelling cultural ecosystem services.

Glycemic control in intensive care patients is complex in terms of patients’ response to care and treatment. The variability and the search for improved insulin therapy outcomes have led to the use of human physiology model based on per-patient metabolic condition to provide personalized automated recommendations. One of the most promising solutions for this is the STAR protocol, which is based on a clinically validated insulin-nutrition-glucose physiological model. However, this approach does not consider demographical background such as age, weight, height, and ethnicity. This article presents the extension to intensive care personalized solution by integrating per-patient demographical, and upon admission information to intensive care conditions to automate decision support for clinical staff. In this context, a virtual study was conducted on 210 retrospectives intensive care patients’ data. To provide a ground, the integration concept is presented roughly, but the details are given in terms of a proof of concept using Bayesian Network, linking the admission background and performance of the STAR control. The proof of concept shows 71.43% and 73.90% overall inference precision, and reliability, respectively, on the test dataset. With more data, improved Bayesian Network is believed to be reproduced. These results, nevertheless, points at the feasibility of the network to act as an effective classifier using intensive care units data, and glycemic control performance to be the basis of a probabilistic, personalized, and automated decision support in the intensive care units.

Background Recently, mobile devices, such as smartphones, have been introduced into healthcare research to substitute paper diaries as data-collection tools in the home environment. Such devices support collecting patient data at different time points over a long period, resulting in clinical time-series data with high temporal complexity, such as time irregularities. Analysis of such time series poses new challenges for machine-learning techniques. The clinical context for the research discussed in this paper is home monitoring in chronic obstructive pulmonary disease (COPD). Objective The goal of the present research is to find out which properties of temporal Bayesian network models allow to cope best with irregularly spaced multivariate clinical time-series data. Methods Two mainstream temporal Bayesian network models of multivariate clinical time series are studied: dynamic Bayesian networks, where the system is described as a snapshot at discrete time points, and continuous time Bayesian networks, where transitions between states are modeled in continuous time. Their capability of learning from clinical time series that vary in nature are extensively studied. In order to compare the two temporal Bayesian network types for regularly and irregularly spaced time-series data, three typical ways of observing time-series data were investigated: (1) regularly spaced in time with a fixed rate; (2) irregularly spaced and missing completely at random at discrete time points; (3) irregularly spaced and missing at random at discrete time points. In addition, similar experiments were carried out using real-world COPD patient data where observations are unevenly spaced. Results For regularly spaced time series, the dynamic Bayesian network models outperform the continuous time Bayesian networks. Similarly, if the data is missing completely at random, discrete-time models outperform continuous time models in most situations. For more realistic settings where data is not missing completely at random, the situation is more complicated. In simulation experiments, both models perform similarly if there is strong prior knowledge available about the missing data distribution. Otherwise, continuous time Bayesian networks perform better. In experiments with unevenly spaced real-world data, we surprisingly found that a dynamic Bayesian network where time is ignored performs similar to a continuous time Bayesian network. Conclusion The results confirm conventional wisdom that discrete-time Bayesian networks are appropriate when learning from regularly spaced clinical time series. Similarly, we found that time series where the missingness occurs completely at random, dynamic Bayesian networks are an appropriate choice. However, for complex clinical time-series data that motivated this research, the continuous-time models are at least competitive and sometimes better than their discrete-time counterparts. Furthermore, continuous-time models provide additional benefits of being able to provide more fine-grained predictions than discrete-time models, which will be of practical relevance in clinical applications.

Global food demand will increase over the next few decades, and sustainable agricultural intensification on current cropland may be a preferred option to meet this demand. Mapping cropping intensity with remote sensing data is of great importance for agricultural production, food security, and agricultural sustainability in the context of global climate change. However, there are some challenges in large-scale cropping intensity mapping. First, existing indicators are too coarse, and fine indicators for measuring cropping intensity are lacking. Second, the regional, intra-class variations detected in time-series remote sensing data across vast areas represent environment-related clusters for each cropping intensity level. However, few existing studies have taken into account the intra-class variations caused by varied crop patterns, crop phenology, and geographical differentiation. In this research, we first presented a new definition, a normalized cropping intensity index (CII), to quantify cropping intensity precisely. We then proposed a Bayesian network model fusing prior knowledge (BNPK) to address the issue of intra-class variations when mapping CII over large areas. This method can fuse regional differentiation factors as prior knowledge into the model to reduce the uncertainty. Experiments on five sample areas covering the main grain-producing areas of mainland China proved the effectiveness of the model. Our research proposes the framework of obtain a CII map with both a finer spatial resolution and a fine temporal resolution at a national scale.

We use Object Oriented Bayesian Networks (OOBNs) to analyze complex ties in the equity market and to detect drivers for the Standard & Poor’s 500 (S&P 500) index. To such aim, we consider a vast number of indicators drawn from various investment areas (Value, Growth, Sentiment, Momentum, and Technical Analysis), and, with the aid of OOBNs, we study the role they played along time in influencing the dynamics of the S&P 500. Our results highlight that the centrality of the indicators varies in time, and offer a starting point for further inquiries devoted to combine OOBNs with trading platforms.

The “Leaky noisy-OR” is a common method used to simplify the elicitation of complex conditional probability tables in Bayesian networks involving Boolean variables. It has proven useful for approximating the required relationship in many real-world situations where there are two or more variables that are potential causes of a single effect variable. However, one of the properties of leaky noisy-OR is Conditional Inter-causal Independence (CII). This property means that ‘explaining away‘ behaviour-one of the most powerful benefits of BN inference is not present when the effect variable is observed as false. Yet, for many real-world problems where the leaky noisy-OR has been considered, this behaviour would be expected, meaning that leaky noisy-OR is deficient as an approximation of the required relationship in such cases. There have been previous attempts to adapt noisy-OR to resolve this problem. However, they require too many additional parameters to be elicited. We describe a simple but powerful extension to leaky noisy-OR that requires only a single additional parameter. While it does not solve the CII problem in all cases, it resolves most of the explaining away deficiencies that occur in practice. The problem and solution is illustrated using an example from intelligence analysis.

The smartphone is a common tool in our everyday lives. However, recent research suggests that using the smartphone has both positive and negative consequences. Although there is no agreement on the concept or the term to label it, researchers and clinical practitioners are worried about the negative consequences derived from excessive smartphone usage. This study aims to analyse the relationship between smartphone addiction and experiential avoidance. A sample of 1176 participants (828 women) with ages ranging from 16 to 82 (M = 30.97; SD = 12.05) was used. The SAS-SV scale was used to measure smartphone addiction and the AAQ-II to assess experiential avoidance. To model the relationship between variables, Bayesian inference and Bayesian networks were used. The results show that experiential avoidance and social networks usage are directly related to smartphone addiction. Additionally, the data suggests that sex is playing a mediating role in the observed relationship between these variables. These results are useful for understanding healthy and pathological interaction with smartphones and could be helpful in orienting or planning future psychological interventions to treat smartphone addiction.

After the ban of battery cages in 1988, a welfare control programme for laying hens was developed in Sweden. Its goal was to monitor and ensure that animal welfare was not negatively affected by the new housing systems. The present observational study provides an overview of the current welfare status of commercial layer flocks in Sweden and explores the complexity of welfare aspects by investigating and interpreting the inter-relationships between housing system, production type (i.e. organic or conventional), facilities, management and animal welfare indicators. For this purpose, a machine learning procedure referred to as structure discovery was applied to data collected through the welfare programme during 2010–2014 in 397 flocks housed in 193 different farms. Seventeen variables were fitted to an Additive Bayesian Network model. The optimal model was identified by an exhaustive search of the data iterated across incremental parent limits, accounting for prior knowledge about causality, potential over-dispersion and clustering. The resulting Directed Acyclic Graph shows the inter-relationships among the variables. The animal-based welfare indicators included in this study – flock mortality, feather condition and mite infestation – were indirectly associated with each other. Of these, severe mite infestations were rare (4% of inspected flocks) and mortality was below the acceptable threshold (< 0.6%). Feather condition scored unsatisfactory in 21% of the inspected flocks; however, it seemed to be only associated to the age of the flock, ruling out any direct connection with managerial and housing variables. The environment-based welfare indicators – lighting and air quality – were an issue in 5 and 8% of the flocks, respectively, and showed a complex inter-relationship with several managerial and housing variables leaving room for several options for intervention. Additive Bayesian Network modelling outlined graphically the underlying process that generated the observed data. In contrast to ordinary regression, it aimed at accounting for conditional independency among variables, facilitating causal interpretation.

Logistic regression techniques can be used to restrict the conditional probabilities of a Bayesian network for discrete variables. More specifically, each variable of the network can be modeled through a logistic regression model, in which the parents of the variable define the covariates. When all main effects and interactions between the parent variables are incorporated as covariates, the conditional probabilities are estimated without restrictions, as in a traditional Bayesian network. By incorporating interaction terms up to a specific order only, the number of parameters can be drastically reduced. Furthermore, ordered logistic regression can be used when the categories of a variable are ordered, resulting in even more parsimonious models. Parameters are estimated by a modified junction tree algorithm. The approach is illustrated with the Alarm network.

Prediction and inference on temporal data is very frequently performed using time series data alone. We believe that these tasks could benefit from leveraging the contex- tual metadata associated to time series - such as location, type, etc. Conversely, tasks involving prediction and inference on metadata could benefit from information held within time series. However, there exists no standard way of jointly modeling both time series data and descriptive metadata. Moreover, metadata frequently contains highly correlated or redundant information, and may contain errors and missing values. We first consider the problem of learning the inherent probabilistic graphical structure of metadata as a Bayesian Network. This has two main benefits: (i) once structured as a graphical model, metadata is easier to use in order to improve tasks on temporal data and (ii) the learned model enables inference tasks on metadata alone, such as missing data imputation. However, Bayesian network structure learning is a tremendous mathematical challenge, that involves a NP-Hard optimization problem. We present a tailor-made structure learning algorithm, inspired from novel theoretical results, that exploits (quasi)-determinist dependencies that are typically present in descriptive metadata. This algorithm is tested on numerous benchmark datasets and some industrial metadatasets containing deterministic relationships. In both cases it proved to be significantly faster than state of the art, and even found more performant structures on industrial data. Moreover, learned Bayesian networks are consistently sparser and therefore more readable. We then focus on designing a model that includes both static (meta)data and dynamic data. Taking inspiration from state of the art probabilistic graphical models for tem- poral data (Dynamic Bayesian Networks) and from our previously described approach for metadata modeling, we present a general methodology to jointly model metadata and temporal data as a hybrid static-dynamic Bayesian network. We propose two main algorithms associated to this representation: (i) a learning algorithm, which while being optimized for industrial data, still generalizes to any task of static and dynamic data fusion, and (ii) an inference algorithm, enabling both usual tasks on temporal or static data alone, and tasks using the two types of data. Finally, we discuss some of the notions introduced during the thesis, including ways to measure the generalization performance of a Bayesian network by a score inspired from the cross-validation procedure from supervised machine learning. We also propose various extensions to the algorithms and theoretical results presented in the previous chapters, and formulate some research perspectives.

Structure learning is one of the main concerns in studies of Bayesian networks. In the present paper, we consider networks consisting of both observable and hidden nodes, and propose a method to investigate the existence of a hidden node between observable nodes, where all nodes are discrete. This corresponds to the model selection problem between the networks with and without the middle hidden node. When the network includes a hidden node, it has been known that there are singularities in the parameter space, and the Fisher information matrix is not positive definite. Then, the many conventional criteria for structure learning based on the Laplace approximation do not work. The proposed method is based on Bayesian clustering, and its asymptotic property justifies the result; the redundant labels are eliminated and the simplest structure is detected even if there are singularities.

Habitat management is increasingly considered as a promising approach to favor the ecosystem service of biological control by enhancing natural enemies. However, habitat management, whether at local or landscape scale, remains very uncertain for farmers. Interactions between ecological processes and agricultural practices are indeed uncertain and site-specific, which makes implementation difficult. Thus, prospecting innovations based on habitat management may benefit from integrating local stakeholders and their knowledge. Our objective is to explore with both local and scientific stakeholders how they perceive agricultural practices, ecological processes, and services related to biological pest control and habitat management. We conducted a participatory Bayesian Network modeling approach with five stakeholders in Southwest France around apple orchard cultivation. We co-constructed such Bayesian Networks based on participants’ knowledge. We explored scenarios favoring natural enemies and habitat manipulation with each participant’s Bayesian Network. We compared how different stakeholders perceive the impact of each scenario on the biological control ecosystem service. Our results indicate that a landscape with a high proportion of semi-natural habitats does not translate into significant biological control for most participants even though some stakeholders perceive a significant impact on generalist predators’ activity within orchards. For these local stakeholders, habitat management at the orchard level such as inter-row vegetation seems currently more promising than at the landscape scale. Here, we show for the first time that the use of Bayesian modeling in a participatory manner can give precious insights into the most promising perspectives on habitat management at different scales. These different local perspectives suggest in particular that further dialogue between ecologists and local stakeholders should be sought about inter-row habitat management as the most promising practice to foster biological pest control and other ecosystem services.

Multivariate models with parsimonious dependence have been used for a large number of variables, and have mainly been developed for multivariate Gaussian. Graphical dependence model representations include Bayesian networks, conditional independence graphs, and truncated vines. The class of Gaussian truncated vines is a subset of Gaussian Bayesian networks and Gaussian conditional independence graphs, but has an extension to non‐Gaussian dependence with (i) combinations of continuous and discrete random variables with arbitrary univariate margins, and (ii) accommodation of latent variables. To illustrate the importance of graphical models with latent variables that do not rely on the Gaussian assumption, the combined factor‐vine structure is presented and applied to a data set of stock returns.

This paper presents an application of the Bayesian belief network for coastal erosion management at the regional scale. A “Bayesian ERosion Management Network” (BERM-N) is developed and trained based on yearly cross-shore profile data available along the Holland coast. Profiles collected for over 50 years and at 604 locations were combined with information on different sand nourishment types (i.e., beach, dune, and shoreface) and volumes implemented during the analyzed time period. The network was used to assess the effectiveness of nourishments in mitigating coastal erosion. The effectiveness of nourishments was verified using two coastal state indicators, namely the momentary coastline position and the dune foot position. The network shows how the current nourishment policy is effective in mitigating the past erosive trends. While the effect of beach nourishment was immediately visible after implementation, the effect of shoreface nourishment reached its maximum only 5–10 years after implementation of the nourishments. The network can also be used as a predictive tool to estimate the required nourishment volume in order to achieve a predefined coastal erosion management objective. The network is interactive and flexible and can be trained with any data type derived from measurements as well as numerical models.

Worldwide the amount of data generated by the medical community is staggering, and increasing dramatically. Using this data to improve patient care using analytics and machine learning is a huge and largely untapped opportunity. The most important medical data captured exist in patients' electronic health records (EHRs) which are maintained and utilized by health care providers. EHRs consist of rich and comprehensive patient-specific information from a large number of sources in different formats with heterogeneous data types. There are numerous challenges in attempting to apply existing analytic tools and methodologies to this data. Many features extracted from EHRs have dependent relationships - for example, “flu” and “high body temperature”. Bayesian networks, as one of the few modeling methodologies which capture feature dependence rather than assuming independence, provide a flexible foundation for modeling EHRs. However, existing Bayesian network learning methodologies produce models whose complexity makes them difficult for clinicians to utilize or even interpret. Therefore, better model visualization methodologies, as well as learning methods which produce models more amenable to simplification and summarization, are critical to making them interpretable and useful to clinicians, and therefore to improving patient care. In this dissertation, I present a framework for predictive analysis of patient clinical data, from feature extraction to model analysis. I first study straightforward machine learning approaches on extracted EHR features and find that incorporating diagnosis features improves area under ROC curve (AUC) by 10% compared to a baseline. Because of the many dependencies between features extracted from EHRs, I next investigate Bayesian network models, in which my clinician collaborators have identified known and suspected high pressure ulcer risk factors. The models also substantially increase sensitivity of the prediction - nearly three times higher comparing to logistical regression models - without sacrificing overall accuracy. However, interpreting these models involves a significant cognitive burden, motivating my investigation of visual analytic techniques. To this end, I develop an interactive tool for visualizing Bayesian networks to improve clinicians’ insight and interpretation of models. I perform a user study to assess the impact of the tool and its features. The results show quantitatively that users complete tasks more efficiently when using the tool, and qualitatively that they found it useful. Bayesian networks containing natural groupings or “clusters” are better suited to visualization and summarization. Since existing Bayesian network learning methods do not naturally yield such groupings, I alter the Bayesian network learning process to learn structures which optimize not just for representing dependency relationships, but additionally and simultaneously, for clusterability measures. My results show that the augmented Bayesian network process can find structures with much larger clusterability measures, with only a small decrease in their standard scoring measure. Visualizations of learned clustered Bayesian networks show that the algorithm cohesively groups related features, making the networks easier to interpret.