Knowledge discovery from observational data for process control using causal Bayesian networks

This paper investigates learning causal relationships from the extensive datasets that are becoming increasingly available in manufacturing systems. A causal modeling approach is proposed to improve an existing causal discovery algorithm by integrating manufacturing domain knowledge with the algorithm. The approach is demonstrated by discovering the causal relationships among the product quality and process variables in a rolling process. When allied with engineering interpretations, the results can be used to facilitate rolling process control.     

贝叶斯网络推理在信息系统安全风险评估中的应用

提出了一种基于贝叶斯网络推理的安全风险评估方法。从实际出发建立信息系统的贝叶斯网络模型，根据专家给出的先验信息，结合获得的证据信息，运用Pearl方法完成对模型的评估，给出在特定条件下模型的计算——线性推理算法。最后，以实例分析信息系统安全评估的实现过程，结果表明，该方法可行、有效。     

Modeling of latent structure of indomethacin solid dispersion tablet using Bayesian networks

Background: When designing pharmaceutical products, the relationships between causal factors and pharmaceutical responses are intricate. A Bayesian network (BN) was used to clarify the latent structure underlying the causal factors and pharmaceutical responses of a tablet containing solid dispersion (SD) of indomethacin (IMC).

Method: IMC, a poorly water-soluble drug, was tested with polyvinylpyrrolidone as the carrier polymer. Tablets containing a SD or a physical mixture of IMC, different quantities of magnesium stearate, microcrystalline cellulose, and low-substituted hydroxypropyl cellulose, and subjected to different compression force were selected as the causal factors. The pharmaceutical responses were the dissolution properties and tensile strength before and after the accelerated test and a similarity factor, which was used as an index of the storage stability.

Result: BN models were constructed based on three measurement criteria for the appropriateness of the graph structure. Of these, the BN model based on Akaike’s information criterion was similar to the results for the analysis of variance. To quantitatively estimate the causal relationships underlying the latent structure in this system, conditional probability distributions were inferred from the BN model. The responses were accurately predicted using the BN model, as reflected in the high correlation coefficients in a leave-one-out cross-validation procedure.

Conclusion: The BN technique provides a better understanding of the latent structure underlying causal factors and responses.     

Bayesian random effect models incorporating real-time weather and traffic data to investigate mountainous freeway hazardous factors

Freeway crash occurrences are highly influenced by geometric characteristics, traffic status, weather conditions and drivers’ behavior. For a mountainous freeway which suffers from adverse weather conditions, it is critical to incorporate real-time weather information and traffic data in the crash frequency study. In this paper, a Bayesian inference method was employed to model one year's crash data on I-70 in the state of Colorado. Real-time weather and traffic variables, along with geometric characteristics variables were evaluated in the models. Two scenarios were considered in this study, one seasonal and one crash type based case. For the methodology part, the Poisson model and two random effect models with a Bayesian inference method were employed and compared in this study. Deviance Information Criterion (DIC) was utilized as a comparison factor. The correlated random effect models outperformed the others. The results indicate that the weather condition variables, especially precipitation, play a key role in the crash occurrence models. The conclusions imply that different active traffic management strategies should be designed based on seasons, and single-vehicle crashes have different crash mechanism compared to multi-vehicle crashes.     

On the use of the hybrid causal logic method in offshore risk analysis

In the Norwegian offshore oil and gas industry risk analyses have been used to provide decision support for more than 20 years. The focus has traditionally been on the planning phase, but during the last years a need for better risk analysis methods for the operational phase has been identified. Such methods should take human and organizational factors into consideration in a more explicit way than the traditional risk analysis methods do. Recently, a framework, called hybrid causal logic (HCL), has been developed based on traditional risk analysis tools combined with Bayesian belief networks (BBNs), using the aviation industry as a case. This paper reviews this framework and discusses its applicability for the offshore industry, and the relationship to existing research projects, such as the barrier and operational risk analysis project (BORA). The paper also addresses specific features of the framework and suggests a new approach for the probability assignment process. This approach simplifies the assignment process considerably without loosing the flexibility that is needed to properly reflect the phenomena being studied.     

Application of Bayesian network to the probabilistic risk assessment of nuclear waste disposal

The scenario in a risk analysis can be defined as the propagating feature of specific initiating event which can go to a wide range of undesirable consequences. If we take various scenarios into consideration, the risk analysis becomes more complex than do without them. A lot of risk analyses have been performed to actually estimate a risk profile under both uncertain future states of hazard sources and undesirable scenarios. Unfortunately, in case of considering specific systems such as a radioactive waste disposal facility, since the behaviour of future scenarios is hardly predicted without special reasoning process, we cannot estimate their risk only with a traditional risk analysis methodology. Moreover, we believe that the sources of uncertainty at future states can be reduced pertinently by setting up dependency relationships interrelating geological, hydrological, and ecological aspects of the site with all the scenarios. It is then required current methodology of uncertainty analysis of the waste disposal facility be revisited under this belief.In order to consider the effects predicting from an evolution of environmental conditions of waste disposal facilities, this paper proposes a quantitative assessment framework integrating the inference process of Bayesian network to the traditional probabilistic risk analysis. We developed and verified an approximate probabilistic inference program for the specific Bayesian network using a bounded-variance likelihood weighting algorithm. Ultimately, specific models, including a model for uncertainty propagation of relevant parameters were developed with a comparison of variable-specific effects due to the occurrence of diverse altered evolution scenarios (AESs). After providing supporting information to get a variety of quantitative expectations about the dependency relationship between domain variables and AESs, we could connect the results of probabilistic inference from the Bayesian network with the consequence evaluation model addressed. We got a number of practical results to improve current knowledge base for the prioritization of future risk-dominant variables in an actual site.     

A Bayesian Belief Network modelling of organisational factors in risk analysis: A case study in maritime transportation

The paper presents an innovative approach to integrate Human and Organisational Factors (HOF) into risk analysis. The approach has been developed and applied to a case study in the maritime industry, but it can also be utilised in other sectors. A Bayesian Belief Network (BBN) has been developed to model the Maritime Transport System (MTS), by taking into account its different actors (i.e., ship-owner, shipyard, port and regulator) and their mutual influences. The latter have been modelled by means of a set of dependent variables whose combinations express the relevant functions performed by each actor. The BBN model of the MTS has been used in a case study for the quantification of HOF in the risk analysis carried out at the preliminary design stage of High Speed Craft (HSC). The study has focused on a collision in open sea hazard carried out by means of an original method of integration of a Fault Tree Analysis (FTA) of technical elements with a BBN model of the influences of organisational functions and regulations, as suggested by the International Maritime Organisation's (IMO) Guidelines for Formal Safety Assessment (FSA). The approach has allowed the identification of probabilistic correlations between the basic events of a collision accident and the BBN model of the operational and organisational conditions. The linkage can be exploited in different ways, especially to support identification and evaluation of risk control options also at the organisational level. Conditional probabilities for the BBN have been estimated by means of experts’ judgments, collected from an international panel of different European countries. Finally, a sensitivity analysis has been carried out over the model to identify configurations of the MTS leading to a significant reduction of accident probability during the operation of the HSC.     

Broiler chickens can benefit from machine learning: support vector machine analysis of observational epidemiological data

Machine-learning algorithms pervade our daily lives. In epidemiology, supervised machine learning has the potential for classification, diagnosis and risk factor identification. Here, we report the use of support vector machine learning to identify the features associated with hock burn on commercial broiler farms, using routinely collected farm management data. These data lend themselves to analysis using machine-learning techniques. Hock burn, dermatitis of the skin over the hock, is an important indicator of broiler health and welfare. Remarkably, this classifier can predict the occurrence of high hock burn prevalence with accuracy of 0.78 on unseen data, as measured by the area under the receiver operating characteristic curve. We also compare the results with those obtained by standard multi-variable logistic regression and suggest that this technique provides new insights into the data. This novel application of a machine-learning algorithm, embedded in poultry management systems could offer significant improvements in broiler health and welfare worldwide.     

Safety analysis in process facilities: Comparison of fault tree and Bayesian network approaches

Safety analysis in gas process facilities is necessary to prevent unwanted events that may cause catastrophic accidents. Accident scenario analysis with probability updating is the key to dynamic safety analysis. Although conventional failure assessment techniques such as fault tree (FT) have been used effectively for this purpose, they suffer severe limitations of static structure and uncertainty handling, which are of great significance in process safety analysis. Bayesian network (BN) is an alternative technique with ample potential for application in safety analysis. BNs have a strong similarity to FTs in many respects; however, the distinct advantages making them more suitable than FTs are their ability in explicitly representing the dependencies of events, updating probabilities, and coping with uncertainties. The objective of this paper is to demonstrate the application of BNs in safety analysis of process systems. The first part of the paper shows those modeling aspects that are common between FT and BN, giving preference to BN due to its ability to update probabilities. The second part is devoted to various modeling features of BN, helping to incorporate multi-state variables, dependent failures, functional uncertainty, and expert opinion which are frequently encountered in safety analysis, but cannot be considered by FT. The paper concludes that BN is a superior technique in safety analysis because of its flexible structure, allowing it to fit a wide variety of accident scenarios.     

Modeling and analysis of automobile warranty data in presence of bias due to customer-rush near warranty expiration limit

Automobile users experiencing soft failures, often delay reporting of warranty claims till the coverage is about to expire. This results into a customer-rush near the warranty expiration limit leading to an occurrence of ‘spikes’ in warranty claims towards the end of warranty period and thereby introducing a bias into the dataset. At the same time, an occurrence of manufacturing/assembly defects in addition to the usage related failures, lead to ‘spikes’ in warranty claims near the beginning of the warranty period. When such data are used to capture the field failures for obtaining feedback on product quality/reliability, it may lead product or reliability engineers to potentially obtain a distorted picture of the reality. Although in reliability studies from automobile warranty data, several authors have addressed the well-recognized issues of incomplete and unclean nature of warranty data, the issue of ‘spikes’ has not received much attention. In this article, we address the issue of ‘spikes’ in the presence of the incomplete and unclean nature of warranty data and provide a methodology to arrive at component-level empirical hazard plots from such automobile warranty data.     

The analysis of ordered categorical data: An overview and a survey of recent developments

This article review methodologies used for analyzing ordered categorical (ordinal) response variables. We begin by surveying models for data with a single ordinal response variable. We also survey recently proposed strategies for modeling ordinal response variables when the data have some type of clustering or when repeated measurement occurs at various occasions for each subject, such as in longitudinal studies. Primary models in that case includemarginal models andcluster-specific (conditional) models for which effects apply conditionally at the cluster level. Related discussion refers to multi-level and transitional models. The main emphasis is on maximum likelihood inference, although we indicate certain models (e.g., marginal models, multi-level models) for which this can be computationally difficult. The Bayesian approach has also received considerable attention for categorical data in the past decade, and we survey recent Bayesian approaches to modeling ordinal response variables. Alternative, non-model-based, approaches are also available for certain types of inference. This work was partially supported by a grant for A. Agresti from NSF and by a research study leave grant from Victoria University for I. Liu.     

Optimization of segmented constrained layer damping with mathematical programming using strain energy analysis and modal data

A new method for enhancement of damping capabilities of segmented constrained layer damping material is proposed. Constrained layer damping has been extensively used since many years to damp flexural vibrations. The shear deformation occurring in the viscoelastic core is mainly responsible for the dissipation of energy. Cutting both the constraining and the constrained layer, which leads to segmentation, increases the shear deformation at that position. This phenomenon is called edge effect. A two-dimensional model of a cantilever beam has been realized for further investigations. An optimization algorithm using mathematical programming is developed in order to identify a cuts arrangement that optimizes the loss factor. The damping efficiency is estimated using the modal strain energy method. The Nelder–Mead simplex method is used to find the best distribution of cuts. In order to take into account geometrical limitations, the exterior point penalty method is used to transform the constrained objective function into an unconstrained objective function. As the optimization problem is not convex, a modal analysis is performed at each mode in order to identify initial cuts positions that lead to a global minimum. Over a large frequency range, the algorithm is able to identify a distribution of cuts that optimizes the loss factor of each mode under consideration.     

ARAMIS project: a more explicit demonstration of risk control through the use of bow-tie diagrams and the evaluation of safety barrier performance

Over the last two decades a growing interest for risk analysis has been noted in the industries. The ARAMIS project has defined a methodology for risk assessment. This methodology has been built to help the industrialist to demonstrate that they have a sufficient risk control on their site.

Risk analysis consists first in the identification of all the major accidents, assuming that safety functions in place are inefficient. This step of identification of the major accidents uses bow–tie diagrams. Secondly, the safety barriers really implemented on the site are taken into account. The barriers are identified on the bow–ties. An evaluation of their performance (response time, efficiency, and level of confidence) is performed to validate that they are relevant for the expected safety function. At last, the evaluation of their probability of failure enables to assess the frequency of occurrence of the accident. The demonstration of the risk control based on a couple gravity/frequency of occurrence is also possible for all the accident scenarios.

During the risk analysis, a practical tool called risk graph is used to assess if the number and the reliability of the safety functions for a given cause are sufficient to reach a good risk control.     

Release of BSA from porous matrices constituted of alginate–Ca and PNIPAAm-interpenetrated networks

The synthesis of thermosensitive Interpenetrating Polymer Network (IPN) hydrogels and the release of Bovine Serum Albumin (BSA) from the hydrogels were reported. The hydrogels, constituted of poly(N-isopropyl acrylamide) PNIPAAm network interpenetrated in alginate–Ca²⁺ network, were synthesized in a two-stepped process. In the first step, PNIPAAm network was synthesized from an aqueous solution containing N-isopropyl acrylamide (NIPAAm) monomers and N,N′-methylene-bis-acrylamide (MBAAm) co-monomers, and sodium alginate (SA) (1 or 2% w/v). The concentration of NIPAAm monomers in the hydrogel-forming solution was always 2.5, 5.0 or 10.0% (w/v). In the second step, alginate–Ca²⁺ networks were formed by immersion of the membrane, obtained on the first step, in a 1.0% (w/v) aqueous calcium chloride. The IPN hydrogels were characterized as a function of temperature (from 25 to 45 °C) through the following measurements: drop water contact angle (DWCA), compression elastic modulus (E) and cross-linking density (ν_e). The morphology was investigated using scanning electronic microscopy (SEM). In vitro release of BSA from the hydrogels was monitored by UV–Vis spectroscopy at 22 °C and 37 °C. DWCA results showed a decrease in the hydrogel hydrophilicity when the temperature and/or the PNIPAAm amount on hydrogels were increased. PNIPAAm-loader hydrogels are more compacted and presented elevated rigidity, mainly above 35 °C. This trend was attributed to the collapsing of PNIPAAm chains as the hydrogels were warmed above its Lower Critical Solution Temperature (LCST), which in aqueous solution is ca. 32–33 °C. The amount of BSA released from the alginate–Ca²⁺/PNIPAAm hydrogels changes inversely to both amount of PNIPAAm and temperature. The transport of BSA from the hydrogels was evaluated through a conventional model. In the lesser-compacted hydrogels the release occurs mostly by diffusion. In the more compacted ones the chain relaxation contributes to the BSA release. Thus, the alginate–Ca²⁺/PNIPAAm IPN-typed matrixes may be considered as smart hydrogels for the release of BSA, because the amount and rate of BSA released may be tailored by both the NIPAAm concentration in the hydrogel-forming solution and the control of temperature of hydrogel.     

Group inference of default‐mode networks from functional magnetic resonance imaging data: comparison of random‐ and mixed‐effects group statistics

Default‐mode network (DMN) activity measured with functional magnetic resonance imaging (fMRI) represents dominant intrinsic neuronal activations of the human brain during rest as opposed to task periods. Previous studies have demonstrated the utility of DMNs in identifying characteristic traits such as hyperactivation and hypoactivation from group‐level fMRI data. However, these group‐level spatial patterns (SPs) were mostly based on random‐effect (RFX) statistics determined using only the intersubject variability. To reduce the potentially significant level of variability in group‐level SPs in RFX due to intrasubject variability, we were motivated to adopt a mixed‐effects (MFX) statistics that is using both intrasubject and intersubject variability. Publicly available group fMRI database during resting state was analyzed using a temporal concatenation‐based group independent component (IC) analysis, and DMN‐related ICs at the group‐level were automatically selected. The individual‐level SPs of these DMN‐related ICs were subsequently estimated using a dual‐regression approach. Using these individual‐level SPs, we evaluated the reproducibility and potential variability of the DMNs from the RFX and MFX statistics using performance measures including (1) neuronal activation levels, (2) percentages of overlap, (3) Pearson's spatial correlation coefficients, and (4) the distances between center‐of‐clusters. The resulting SPs from the MFX‐based group inference showed a significantly greater level of reproducibility than those from the RFX‐based group inference as tested in a bootstrapping framework Family‐wise error (FWE)‐corrected p < 10^?10, one‐way analysis of variance (ANOVA)). The reported findings may provide a valuable supplemental option for investigating the neuropsychiatric group‐ or condition‐dependent characteristic traits implicated in DMNs. © 2012 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 22, 121–131, 2012