Real-Time and Data-Driven Operation Optimization and Knowledge Discovery for an Enterprise Information System

An enterprise information system (EIS) is an integrated data-applications platform characterized by diverse, heterogeneous, and distributed data sources. For many enterprises, a number of business processes still depend heavily on static rule-based methods and extensive human expertise. Enterprises are faced with the need for optimizing operation scheduling, improving resource utilization, discovering useful knowledge, and making data-driven decisions.

This thesis research is focused on real-time optimization and knowledge discovery that addresses workflow optimization, resource allocation, as well as data-driven predictions of process-execution times, order fulfillment, and enterprise service-level performance. In contrast to prior work on data analytics techniques for enterprise performance optimization, the emphasis here is on realizing scalable and real-time enterprise intelligence based on a combination of heterogeneous system simulation, combinatorial optimization, machine-learning algorithms, and statistical methods.

On-demand digital-print service is a representative enterprise requiring a powerful EIS.We use real-life data from Reischling Press, Inc. (RPI), a digit-print-service provider (PSP), to evaluate our optimization algorithms.

In order to handle the increase in volume and diversity of demands, we first present a high-performance, scalable, and real-time production scheduling algorithm for production automation based on an incremental genetic algorithm (IGA). The objective of this algorithm is to optimize the order dispatching sequence and balance resource utilization. Compared to prior work, this solution is scalable for a high volume of orders and it provides fast scheduling solutions for orders that require complex fulfillment procedures. Experimental results highlight its potential benefit in reducing production inefficiencies and enhancing the productivity of an enterprise.

We next discuss analysis and prediction of different attributes involved in hierarchical components of an enterprise. We start from a study of the fundamental processes related to real-time prediction. Our process-execution time and process status prediction models integrate statistical methods with machine-learning algorithms. In addition to improved prediction accuracy compared to stand-alone machine-learning algorithms, it also performs a probabilistic estimation of the predicted status. An order generally consists of multiple series and parallel processes. We next introduce an order-fulfillment prediction model that combines advantages of multiple classification models by incorporating flexible decision-integration mechanisms. Experimental results show that adopting due dates recommended by the model can significantly reduce enterprise late-delivery ratio. Finally, we investigate service-level attributes that reflect the overall performance of an enterprise. We analyze and decompose time-series data into different components according to their hierarchical periodic nature, perform correlation analysis,

and develop univariate prediction models for each component as well as multivariate models for correlated components. Predictions for the original time series are aggregated from the predictions of its components. In addition to a significant increase in mid-term prediction accuracy, this distributed modeling strategy also improves short-term time-series prediction accuracy.

In summary, this thesis research has led to a set of characterization, optimization, and prediction tools for an EIS to derive insightful knowledge from data and use them as guidance for production management. It is expected to provide solutions for enterprises to increase reconfigurability, accomplish more automated procedures, and obtain data-driven recommendations or effective decisions.

Growth and production of Pacific ocean perch (Sebastes alutus) in nursery habitats of the Gulf of Alaska

Nursery areas for juvenile fishes are often important for determining recruitment in marine populations by providing habitats that can maximize growth and thereby minimize mortality. Pacific ocean perch (POP, Sebastes alutus) have an extended juvenile period where they inhabit rocky nursery habitats. We examined POP nursery areas to link growth potential to recruitment. Juvenile POP were captured from nursery areas in 2004 and 2008, and estimated growth rates ranged from −0.19 to 0.60 g day−1 based on differences in size between June and August. Predicted growth rates from a bioenergetics model ranged from 0.05 to 0.49 g day−1 and were not significantly different than observed. Substrate preferences and the distribution of their preferred habitats were utilized to predict the extent of juvenile POP nursery habitat in the Gulf of Alaska. Based on densities of fish observed on underwater video transects and the spatial extent of nursery areas, we predicted 278 and 290 million juvenile POP were produced in 2004 and 2008. Growth potential for juvenile POP was reconstructed using the bioenergetics model, spring zooplankton bloom timing and duration and bottom water temperature for 1982–2008. When a single outlying recruitment year in 1986 was removed, growth potential experienced by juvenile POP in nursery areas was significantly correlated to the recruitment time-series from the stock assessment, explaining ∼30% of the variability. This research highlights the potential to predict recruitment using habitat-based methods and provides a potential mechanism for explaining some of the POP recruitment variability observed for this population.

Dynamics of air–sea CO2 fluxes in the northwestern European shelf based on voluntary observing ship and satellite observations

From January 2011 to December 2013, we constructed a comprehensive pCO2 data set based on voluntary observing ship (VOS) measurements in the western English Channel (WEC). We subsequently estimated surface pCO2 and air–sea CO2 fluxes in northwestern European continental shelf waters using multiple linear regressions (MLRs) from remotely sensed sea surface temperature (SST), chlorophyll a concentration (Chl a), wind speed (WND), photosynthetically active radiation (PAR) and modeled mixed layer depth (MLD). We developed specific MLRs for the seasonally stratified northern WEC (nWEC) and the permanently well-mixed southern WEC (sWEC) and calculated surface pCO2 with uncertainties of 17 and 16 μatm, respectively. We extrapolated the relationships obtained for the WEC based on the 2011–2013 data set (1) temporally over a decade and (2) spatially in the adjacent Celtic and Irish seas (CS and IS), two regions which exhibit hydrographical and biogeochemical characteristics similar to those of WEC waters. We validated these extrapolations with pCO2 data from the SOCAT and LDEO databases and obtained good agreement between modeled and observed data. On an annual scale, seasonally stratified systems acted as a sink of CO2 from the atmosphere of −0.6 ± 0.3, −0.9 ± 0.3 and −0.5 ± 0.3 mol C m−2 yr−1 in the northern Celtic Sea, southern Celtic sea and nWEC, respectively, whereas permanently well-mixed systems acted as source of CO2 to the atmosphere of 0.2 ± 0.2 and 0.3 ± 0.2 mol C m−2 yr−1 in the sWEC and IS, respectively. Air–sea CO2 fluxes showed important inter-annual variability resulting in significant differences in the intensity and/or direction of annual fluxes. We scaled the mean annual fluxes over these provinces for the last decade and obtained the first annual average uptake of −1.11 ± 0.32 Tg C yr−1 for this part of the northwestern European continental shelf. Our study showed that combining VOS data with satellite observations can be a powerful tool to estimate and extrapolate air–sea CO2 fluxes in sparsely sampled area.

Self-reported quality of life of 8-12 year old children with cerebral palsy: a cross-sectional European Study

Background

Little is known about the quality of life (QoL) of disabled children. We describe self-reported QoL of children with cerebral palsy, factors that influence it, and how it compares with QoL of the general population.

Methods

1174 children aged 8–12 years were randomly selected from eight population-based registers of children with cerebral palsy in six European countries and 743 (63%) agreed to participate; one further region recruited 75 children from multiple sources. Researchers visited these 818 children. 318 (39%) with severe intellectual impairment could not self-report; 500 (61%) reported their QoL using KIDSCREEN, an instrument with scores in ten domains, each with SD=10. Multivariable regression was used to relate QoL to impairments, pain, and sociodemographic characteristics. Comparisons were made with QoL data from the general population.

Findings

Impairments were not significantly associated with six KIDSCREEN domains. Comparison of least and most able groups showed that severely limited self-mobility was significantly associated with reduced mean score for physical wellbeing (7·6, 95% CI 2·7–12·4); intellectual impairment with reduced mean for moods and emotions (3·7, 1·5–5·9) and autonomy (3·3, 0·9–5·7); and speech difficulties with reduced mean for relationships with parents (4·5, 1·9–7·1). Pain was common and associated with lower QoL on all domains. Impairments and pain explained up to 3% and 7%, respectively, of variation in QoL. Children with cerebral palsy had similar QoL to children in the general population in all domains except schooling, in which evidence was equivocal, and physical wellbeing, in which comparison was not possible.

Interpretation

Parents can be reassured that most children aged 8–12 years with cerebral palsy will have similar QoL to other children. This finding should guide social and educational policy to ensure that disabled children participate fully in society. Because of its association with QoL, children's pain should be carefully assessed.

Brief review of invariant texture analysis methods

This paper considers invariant texture analysis. Texture analysis approaches whose performances are not affected by translation, rotation, affine, and perspective transform are addressed. Existing invariant texture analysis algorithms are carefully studied and classified into three categories: statistical methods, model based methods, and structural methods. The importance of invariant texture analysis is presented first. Each approach is reviewed according to its classification, and its merits and drawbacks are outlined. The focus of possible future work is also suggested.

A prospective clinical trial of a real-time polymerase chain reaction assay for the diagnosis of candidemia in nonneutropenic, critically ill adults

Background. Invasive Candida infection among nonneutropenic, critically ill adults is a clinical problem that has received increasing attention in recent years. Poor performance of extant diagnostic modalities has promoted risk-based, preemptive prescribing in view of the poor outcomes associated with inadequate or delayed antifungal therapy; this risks unnecessary overtreatment. A rapid, reliable diagnostic test could have a substantial impact on therapeutic practice in this patient population.

Methods. Three TaqMan-based real-time polymerase chain reaction assays were developed that are capable of detecting the main medically important Candida species, categorized according to the likelihood of fluconazole susceptibility. Assay 1 detected Candida albicans, Candida parapsilosis, Candida tropicalis, and Candida dubliniensis. Assays 2 and 3 detected Candida glabrata and Candida krusei, respectively. The clinical performance of these assays, applied to serum, was evaluated in a prospective trial of nonneutropenic adults in a single intensive care unit.

Results. In all, 527 specimens were obtained from 157 participants. All 3 assays were run in parallel for each specimen; they could be completed within 1 working day. Of these, 23 specimens were obtained from 23 participants categorized as having proven Candida infection at the time of sampling. If a single episode of Candida famata candidemia was excluded, the estimated clinical sensitivity, specificity, and positive and negative predictive values of the assays in this trial were 90.9%, 100%, 100% and 99.8%, respectively.

Conclusions. These data suggest that the described assays perform well in this population for enhancing the diagnosis of candidemia. The extent to which they may affect clinical outcomes, prescribing practice, and cost-effectiveness of care remains to be ascertained.

Prediction- and simulation-error based perceptron training: Solution space analysis and a novel combined training scheme

Previous papers have noted the difficulty in obtaining neural models which are stable under simulation when trained using prediction-error-based methods. Here the differences between series-parallel and parallel identification structures for training neural models are investigated. The effect of the error surface shape on training convergence and simulation performance is analysed using a standard algorithm operating in both training modes. A combined series-parallel/parallel training scheme is proposed, aiming to provide a more effective means of obtaining accurate neural simulation models. Simulation examples show the combined scheme is advantageous in circumstances where the solution space is known or suspected to be complex. (c) 2006 Elsevier B.V. All rights reserved.

Conceptions of Teacher Identity: Creating Self-Boxes as Arts-Based Self-Study

This paper presents a palimpsest of ways in which self-study draws upon arts-based methods not just as processes towards teacher development, but also as means to problematize and inquire into conceptualizations of the self. It focuses on the creation of individual self-boxes that mediate teachers’ dynamic narratives of identity. Concepts of the unitary self, the decentred self and the relationship between inner and outer experience are challenged and illustrated through two interlapping stories made manifest through the creation of self-boxes. From time immemorial man has known that he is the subject most deserving of his own study, but he has also fought shy of treating this subject as a whole, that is, in accordance with its total character. (Buber, 1947, p. 140)

Introduction to genetic epidemiology

Genetic epidemiology is of topical and increasingly practical relevance. The subject attempts to answer 2 questions: (1) is there a genetic component to a disease, and (2) what genes are involved? This article summarizes genetic epidemiologic methods, describing family- and population-based methods used to locate and identify genes and the advantages and disadvantages of each approach. Health care professionals are faced with more and more genetic information, both from interested patients and from the media, and understanding the principles underlying genetic studies allows such information to be placed in context. © 2010 American Optometric Association.

Current and possible future methods of assessing the activities of photocatalyst films

The continuing interest in semiconductor photochemistry, SPC, and the emergence of commercial products that utilise films of photocatalyst materials, has created an urgent need to agree a set of methods for assessing photocatalytic activity and international committees are now meeting to address this issue. This article provides a brief overview of two of the most popular current methods employed by researchers for assessing SPC activity. and one which has been published just recently and might gain popularity in the future, given its ease of use. These tests are: the stearic acid (SA) test, the methylene blue (MB) test and the resazurin (Rz) ink test, respectively. The basic photochemical and chemical processes that underpin each of these tests are described, along with typical results for laboratory made sol-gel titania films and a commercial form of self-cleaning glass, Activ (TM). The pros and cons of their future use as possible standard assessment techniques are considered. (C) 2007 Elsevier B.V. All rights reserved.

Association analysis of proopiomelanocortin (POMC) haplotypes in type 1 diabetes in a UK population

Aim

To assess the association of POMC haplotype-tagged single nucleotide polymorphisms (htSNPs) with the development of type 1 diabetes (T1D) in a Caucasian population.

Methods

All exons, intron 1, and approximately 6-kb upstream and 3-kb downstream of the POMC gene were bidirectionally resequenced to identify DNA polymorphisms in 30 individuals. Allele frequencies were determined (60 chromosomes) and efficient htSNPs were selected using the htSNP2 programme. Genotyping was performed in 390 cases, 339 controls and 245 T1D parent-offspring trios, using Taqman, Sequenom and direct-sequencing technologies.

Results

Thirteen polymorphisms (two novel) with a minor allele frequency greater than 1% were identified. Six POMC htSNPs (rs3754863 G>A, ss161151662 A>G, rs3754860 C>T, rs1009388 G>C, rs3769671 A>C, rs1042571 G>A) were identified. Allele and haplotype frequencies were similar between case and control groups (P>0.60 by permutation test), and assessment of allele transmission distortion from informative parents to affected offspring also failed to find any association. Stratification of these analyses for age-at-onset and HLA-DR risk group (DR3/DR4) revealed no significant associations. A haplotype block of 9.86-kb from rs3754863 to rs1042571 was identified, encompassing the POMC gene. Comparison of haplotype frequencies identified the GGCGAG haplotype as protective against T1D in 12.9% of cases vs. 18.3% of controls: ?2=8.18, Pc=0.03 by permutation test.

Conclusion

The POMC SNP haplotype GGCGAG may have a protective effect against T1D in the UK population. However, this finding needs to be replicated, and the cellular and molecular processes influenced by this POMC haplotype determined to fully appreciate its impact.

A posteriori-derived dietary patterns and retinal vessel caliber in an elderly population

Purpose. To examine the association between a posteriori–derived dietary patterns (DP) and retinal vessel caliber in an elderly population.

Methods. This was a cross-sectional study of 288 elderly adults (>65 years) who participated in the European Eye study (EUREYE) Northern Irish cohort. DP were extracted using principal component analysis from completed food frequency questionnaires. Semi-automated computer grading was used to determine the mean retinal vessel diameters (central retinal arteriole equivalent [CRAE] and central retinal venule equivalent [CRVE]) from digitized visual field one images using a standard measurement protocol.

Results. Three major DP were identified in this population, which accounted for 21% of the total variance: a “healthy” pattern with high factor loadings for oily fish, fruits and vegetables, and olive oil; an “unhealthy” pattern with high factor loadings for red and processed meat, refined grains, eggs, butter, sugar and sweets; and a “snack and beverage” pattern with high factor loading for pizza, nuts, and coffee. Multivariable linear regression analysis indicated no significant association between major identified DP and mean CRAE or CRVE in all models.

Conclusions. This is the first study to investigate associations between a posteriori–derived DP and retinal vessel caliber. There was no evidence of a relationship between extracted DP and retinal vessel measurements in this population. However, it is possible that potentially important relationships exist between single nutrients or foods and vessel diameters that cannot be identified using a DP approach. Further studies to examine the role of dietary factors in the microcirculation are required.

High population prevalence of cardiac troponin I measured by a high-sensitivity assay and cardiovascular risk estimation: the MORGAM Biomarker Project Scottish Cohort

Aims
Our aim was to test the prediction and clinical applicability of high-sensitivity assayed troponin I for incident cardiovascular events in a general middle-aged European population.

Methods and results
High-sensitivity assayed troponin I was measured in the Scottish Heart Health Extended Cohort (n = 15 340) with 2171 cardiovascular events (including acute coronary heart disease and probable ischaemic strokes), 714 coronary deaths (25% of all deaths), 1980 myocardial infarctions, and 797 strokes of all kinds during an average of 20 years follow-up. Detection rate above the limit of detection (LoD) was 74.8% in the overall population and 82.6% in men and 67.0% in women. Troponin I assayed by the high-sensitivity method was associated with future cardiovascular risk after full adjustment such as that individuals in the fourth category had 2.5 times the risk compared with those without detectable troponin I (P < 0.0001). These associations remained significant even for those individuals in whom levels of contemporary-sensitivity troponin I measures were not detectable. Addition of troponin I levels to clinical variables led to significant increases in risk prediction with significant improvement of the c-statistic (P < 0.0001) and net reclassification (P < 0.0001). A threshold of 4.7 pg/mL in women and 7.0 pg/mL in men is suggested to detect individuals at high risk for future cardiovascular events.

Conclusion
Troponin I, measured with a high-sensitivity assay, is an independent predictor of cardiovascular events and might support selection of at risk individuals.

Two-Stage Orthogonal Least Squares Methods for Neural Network Construction

A number of neural networks can be formulated as the linear-in-the-parameters models. Training such networks can be transformed to a model selection problem where a compact model is selected from all the candidates using subset selection algorithms. Forward selection methods are popular fast subset selection approaches. However, they may only produce suboptimal models and can be trapped into a local minimum. More recently, a two-stage fast recursive algorithm (TSFRA) combining forward selection and backward model refinement has been proposed to improve the compactness and generalization performance of the model. This paper proposes unified two-stage orthogonal least squares methods instead of the fast recursive-based methods. In contrast to the TSFRA, this paper derives a new simplified relationship between the forward and the backward stages to avoid repetitive computations using the inherent orthogonal properties of the least squares methods. Furthermore, a new term exchanging scheme for backward model refinement is introduced to reduce computational demand. Finally, given the error reduction ratio criterion, effective and efficient forward and backward subset selection procedures are proposed. Extensive examples are presented to demonstrate the improved model compactness constructed by the proposed technique in comparison with some popular methods.

Generating a Word-Emotion Lexicon from #Emotional Tweets

Research in emotion analysis of text suggest that emotion lexicon based features are superior to corpus based n-gram features. However the static nature of the general purpose emotion lexicons make them less suited to social media analysis, where the need to adopt to changes in vocabulary usage and context is crucial. In this paper we propose a set of methods to extract a word-emotion lexicon automatically from an emotion labelled corpus of tweets. Our results confirm that the features derived from these lexicons outperform the standard Bag-of-words features when applied to an emotion classification task. Furthermore, a comparative analysis with both manually crafted lexicons and a state-of-the-art lexicon generated using Point-Wise Mutual Information, show that the lexicons generated from the proposed methods lead to significantly better classi- fication performance.