Additive Bayesian Networks Research Articles

Additive Bayesian Networks

Additive Bayesian Networks

Multivariate variable selection in N-of-1 observational studies via additive Bayesian networks.

An N-of-1 observational design characterizes associations among several variables over time in a single individual. Traditional statistical models recommended for experimental N-of-1 trials may not adequately model these observational relationships. We propose an additive Bayesian network using a generalized linear mixed-effects model for the local mean as a novel method for modeling each of these relationships in a data-driven manner. We validate our approach via simulation studies and apply it to a 12-month observational N-of-1 study exploring the impact of stress on daily exercise engagement. We demonstrate the improved performance of the additive Bayesian network to recover the underlying network structure. From the empirical study, we found statistically discernible associations between reports of stress and physical activity on a population level, but these associations may differ at an individual level.

Additive dynamic Bayesian networks for enhanced feature learning in soft sensor modeling

Due to the advantages of indicating variable structure and efficient reasoning, Bayesian Networks (BN) have been widely used in data-driven soft sensor applications. However, restricted to linear and conditional Gaussian property, BN-based soft sensors rarely achieve high prediction accuracy. In this paper, an ensemble learning framework – additive dynamic Bayesian networks (ADBN) is proposed for enhanced feature learning, in which Dynamic Bayesian networks are used to learn the conditional independencies among variables and construct feature sets for the following base learners. Additional DBNs are constructed upon the residual information from the past model, to carry out feature learning to fit the residuals. The procedure is repeated and a termination rule from the perspective of feature learning is proposed to end this process, and thus the model complexity can be well restricted. The proposed method is validated on two actual industrial cases. It reveals that the ADBN feature learning method has obtained great improvements. Compared to the single DBN feature engineering method, the root mean square error (RMSE) performance has been improved by 20% and 13% on the two cases, respectively.

Machine Learning Analysis of Longevity-Associated Gene Expression Landscapes in Mammals.

One of the important questions in aging research is how differences in transcriptomics are associated with the longevity of various species. Unfortunately, at the level of individual genes, the links between expression in different organs and maximum lifespan (MLS) are yet to be fully understood. Analyses are complicated further by the fact that MLS is highly associated with other confounding factors (metabolic rate, gestation period, body mass, etc.) and that linear models may be limiting. Using gene expression from 41 mammalian species, across five organs, we constructed gene-centric regression models associating gene expression with MLS and other species traits. Additionally, we used SHapley Additive exPlanations and Bayesian networks to investigate the non-linear nature of the interrelations between the genes predicted to be determinants of species MLS. Our results revealed that expression patterns correlate with MLS, some across organs, and others in an organ-specific manner. The combination of methods employed revealed gene signatures formed by only a few genes that are highly predictive towards MLS, which could be used to identify novel longevity regulator candidates in mammals.

Biosecurity risk factors for highly pathogenic avian influenza (H5N8) virus infection in duck farms, France.

Highly pathogenic avian influenza (HPAI) subtype H5N8 outbreaks occurred in poultry farms in France in 2016-2017, resulting in significant economic losses and disruption to the poultry industry. Current evidence on associations between actual on-farm biosecurity risk factors and H5N8 occurrence is limited. Therefore, a retrospective matched case-control study was undertaken to investigate the inter-relationships between on-farm biosecurity practices and H5N8 infection status to provide new insights regarding promising targets for intervention. Data were collected on 133 case and 133 control duck farms (i.e. the most affected species) located in one area of the country that was mostly affected by the disease. Data were analysed using Additive Bayesian Networks which offer a rich modelling framework by graphically illustrating the dependencies between variables. Factors indirectly and directly positively associated with farm infection were inadequate management of vehicle movements (odds ratio [OR] 9.3, 95% credible interval [CI] 4.0-22.8) and inadequate delimitation of farm and units (OR 3.0, 95% CI 1.6-5.8), respectively. Inadequate disposal of dead birds was instead negatively associated with the outcome (OR 0.1, 95% CI 0.0-0.3). The findings highlight that reinforcing farm access control systems and reducing the number of visitors are key biosecurity measures to control farm vulnerability to H5N8 infection and could help setting priorities in biosecurity practices to prevent outbreaks' re-occurrence.

Self-Perceived Health, Objective Health, and Quality of Life among People Aged 50 and Over: Interrelationship among Health Indicators in Italy, Spain, and Greece.

It is well known that self-perceived health (SPH), even if it is a subjective health indicator, is significantly associated with objective health and quality of life (QoL) in the general population. Whether it can be considered an indicator of cognitive functioning and quality of life in the elderly is still an open issue. This study used a data-driven approach to investigate the interrelationship among SPH, non-communicable diseases (NCDs), QoL, and cognitive functioning to answer this question. The study sample included information about 12,831 people living in Italy, Spain, and Greece, extracted from the Survey on Health, Aging, and Retirement in Europe, in the year 2015. The additive Bayesian networks methodology was used to identify the best directed acyclic graphs (DAG) for SPH, QoL, and NCDs. Results were given as posterior estimates of generalized linear models (GLM) coefficients, with 95% credibility intervals. Good SPH was associated with a decreasing number of chronic diseases in Italy (coeff = −0.52, 95%CI: [−0.59, −0.44]), Spain (coeff = −0.53, 95%CI: [−0.60, −0.46]) and Greece (coeff = −0.57, 95%CI: [−0.64, −0.50]). Age and Body Mass Index were determinants of NCDs in all countries. QoL of elderly was associated with SPH in Italy (coeff = 0.12, 95%CI: [0.10, 0.14]), Spain (coeff = 0.16, 95%CI: [0.15, 0.18]), and Greece (coeff = 0.18, 95%CI: [0.16, 0.20]). The number of NCDs was higher for people who were not employed in Spain (coeff = 0.45, 95%CI: [0.37, 0.53]) and was decreasing for a unitary increase in years of education in Greece (coeff = −0.12, 95%CI: [−0.14, −0.09]). As a general rule, the framework of the interrelationship among NCDs, SPH, and QoL was similar for Italy, Spain, and Greece. The connections found among indicators could be proposed to identify strategies for health promotion and healthy aging among people aged 50 and above, which are viable in general and at a country level. Reinforcing strategies targeted at some health indicators could have relevant effects on other related indicators.

Additive Bayesian networks for antimicrobial resistance and potential risk factors in non-typhoidal Salmonella isolates from layer hens in Uganda

BackgroundMulti-drug resistant bacteria are seen increasingly and there are gaps in our understanding of the complexity of antimicrobial resistance, partially due to a lack of appropriate statistical tools. This hampers efficient treatment, precludes determining appropriate intervention points and renders prevention very difficult.MethodsWe re-analysed data from a previous study using additive Bayesian networks. The data contained information on resistances against seven antimicrobials and seven potential risk factors from 86 non-typhoidal Salmonella isolates from laying hens in 46 farms in Uganda.ResultsThe final graph contained 22 links between risk factors and antimicrobial resistances. Solely ampicillin resistance was linked to the vaccinating person and disposal of dead birds. Systematic associations between ampicillin and sulfamethoxazole/trimethoprim and chloramphenicol, which was also linked to sulfamethoxazole/trimethoprim were detected. Sulfamethoxazole/trimethoprim was also directly linked to ciprofloxacin and trimethoprim. Trimethoprim was linked to sulfonamide and ciprofloxacin, which was also linked to sulfonamide. Tetracycline was solely linked to ciprofloxacin.ConclusionsAlthough the results needs to be interpreted with caution due to a small data set, additive Bayesian network analysis allowed a description of a number of associations between the risk factors and antimicrobial resistances investigated.

Revealing the structure of the associations between housing system, facilities, management and welfare of commercial laying hens using Additive Bayesian Networks

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.

Estimation of Obstructive Sleep Apnea Severity Using Additive Bayesian Networks

Estimation of Obstructive Sleep Apnea Severity Using Additive Bayesian Networks

Energy cost minimization in plant factories considering weather factors using additive Bayesian networks

In recent years, plant factories have been drawing attention for their capability to solve global food crisis. However, the energy cost of plant factories is high due to the power consumption of ai...

Attitudes of Austrian veterinarians towards euthanasia in small animal practice: impacts of age and gender on views on euthanasia.

BackgroundEuthanasia of pets has been described by veterinarians as “the best and the worst” of the profession. The most commonly mentioned ethical dilemmas veterinarians face in small animal practice are: limited treatment options due to financial constraints, euthanizing of healthy animals and owners wishing to continue treatment of terminally ill animals. The aim of the study was to gain insight into the attitudes of Austrian veterinarians towards euthanasia of small animals. This included assessing their agreement with euthanasia in exemplified case scenarios, potentially predicted by demographic variables (e.g. gender, age, working in small animal practice, employment, working in a team, numbers of performed euthanasia). Further describing the veterinarians’ agreement with a number of different normative and descriptive statements, including coping strategies. A questionnaire with nine euthanasia scenarios, 26 normative and descriptive statements, and demographic data were sent to all members of the Austrian Chamber of Veterinary Surgeons (n = 2478).ResultsIn total, 486 veterinarians answered sufficiently completely to enable analyses. Responses were first explored descriptively before being formally analysed using linear regression and additive Bayesian networks – a multivariate regression methodology – in order to identify joint relationships between the demographic variables, the statements and each of the nine euthanasia scenarios. Mutual dependencies between the demographic variables were found, i.e. female compared to male veterinarians worked mostly in small animal practice, and working mostly in small animal practice was linked to performing more euthanasia per month.ConclusionsGender and age were found to be associated with views on euthanasia: female veterinarians and veterinarians having worked for less years were more likely to disagree with euthanasia in at least some of the convenience euthanasia scenarios. The number of veterinarians working together was found to be the variable with the highest number of links to other variables, demographic as well as ethical statements. This highlights the role of a team potentially providing support in stressful situations. The results are useful for a better understanding of coping strategies for veterinarians with moral stress due to euthanasia of small animals.Electronic supplementary materialThe online version of this article (doi:10.1186/s12917-016-0649-0) contains supplementary material, which is available to authorized users.

Discovering complex interrelationships between socioeconomic status and health in Europe: A case study applying Bayesian Networks

Studies assume that socioeconomic status determines individuals’ states of health, but how does health determine socioeconomic status? And how does this association vary depending on contextual differences? To answer this question, our study uses an additive Bayesian Networks model to explain the interrelationships between health and socioeconomic determinants using complex and messy data. This model has been used to find the most probable structure in a network to describe the interdependence of these factors in five European welfare state regimes. The advantage of this study is that it offers a specific picture to describe the complex interrelationship between socioeconomic determinants and health, producing a network that is controlled by socio-demographic factors such as gender and age. The present work provides a general framework to describe and understand the complex association between socioeconomic determinants and health.

Identifying associations in Escherichia coli antimicrobial resistance patterns using additive Bayesian networks

While the genesis of antimicrobial resistance (AMR) in animal production is a high profile topic in the media and the scientific community, it is still not well understood. The epidemiology of AMR is complex. This complexity is demonstrated by extensive biological and evolutionary mechanisms which are potentially impacted by farm management and husbandry practices – the risk factors. Many parts of this system have yet to be fully described. Notably, the occurrence of multiple resistance patterns is the rule rather than exception – the multivariate problem. A first essential step in the development of any comprehensive risk factor analysis – whose goal is the prevention or reduction of AMR – is to describe those associations between different patterns of resistance which are systematic. That is, have sufficient statistical support for these patterns to be considered robust features of the underlying epidemiological system, and whose presence must therefore be incorporated into any risk factor analysis of AMR for it to be meaningful with respect to the farm environment. Presented here is a case study that seeks to identify systematic associations between patterns of resistance to 13 different antimicrobials in Escherichia coli isolates obtained from composite finisher (>80kg) pig faecal samples obtained from Canada's five major pork producing provinces. The use of a Bayesian network analysis approach allowed us to identify many systematic associations between individual antimicrobial resistances. Sixteen of these resistances are corroborated with existing literature. These associations are distributed between several important classes of antimicrobials including the β-lactams, folate biosynthesis inhibitors, tetracyclines, aminoglycosides and quinolones. This study presents an exciting first step towards the larger and far more ambitious goal of developing generic and holistic risk factor analyses for on-farm occurrence of AMR. Analyses of this nature would combine multivariate response variables (joint patterns of resistance) with multi-factorial causal factors from within the livestock production environment thereby permitting a more complete understanding of the epidemiology of antimicrobial resistance.