Random Sub-sampling Cross-validation Research Articles

Integration of electromagnetic, resistivity-based and production logging data for validating lithofacies and permeability predictive models with tree ensemble algorithms in heterogeneous carbonate reservoirs

This study develops an innovative workflow to identify discrete lithofacies distributions with respect to the well-log records exploiting two tree-based ensemble learning algorithms: extreme gradient boosting (XGBoost) and adaptive boosting (AdaBoost). In the next step, the predicted discrete lithofacies distribution is further assessed with well-log data using an XGBoost regression to predict reservoir permeability. The input well-logging records are gamma ray, neutron porosity, bulk density, compressional slowness, and deep and shallow resistivity. These data originate from a carbonate reservoir in the Mishrif Basin of southern Iraq's oilfield. To achieve a solid prediction of lithofacies permeability, random subsampling cross-validation was applied to the original dataset to formulate two subsets: training for model tuning and testing for the prediction of subsets that are not observed during the model training. The values for the total correct percentage (TCP) of lithofacies predictions for the entire dataset and testing subset were 98 and 93% using the XGBoost algorithm, and 97 and 89% using the AdaBoost classifier, respectively. The XGBoost predictive models led in attaining the least uncertain lithofacies and permeability records for the cored data. For further validation, the predicted lithofacies and reservoir permeability were then compared with porosity–permeability values derived from the nuclear magnetic resonance (NMR) log, the secondary porosity of the full-bore micro imager (FMI) and the production contribution from the production–logging tool (PLT). Therefore, it is believed that the XGBoost model is capable of making accurate predictions of lithofacies and permeability for the same well's non-cored intervals and other non-cored wells in the investigated reservoir.

A physics-guided eXtreme gradient boosting model for predicting the initial productivity of oil wells

The initial productivity of oil wells is an important index in oilfield development. Due to the nonlinear relationship among the numerous affecting factors of productivity, productivity prediction methods based on seepage theory are difficult to characterize this nonlinear relationship. The machine learning method provides a novel theory to solve this problem. However, the data-driven machine learning method neglects the physical-based seepage theory. For balancing the advantage of the seepage model and the data-driven model, the productivity formula was used to guide the machine learning algorithm for achieving nonlinear regression in this study. Therefore, an innovative physics-guided eXtreme Gradient Boosting (XGBoost) trees algorithm was proposed to predict the initial productivity. The productivity formula was employed as the physical constraint and combined with the mean-square error (MSE) loss function of the XGBoost algorithm. A hyperparameter was introduced to balance the productivity formula and the MSE formula for constructing this adaptive physics-guided loss function. Moreover, the random subsampling cross-validation approach was adopted to train the XGBoost model. The grid search approach was introduced to optimize the hyperparameters of the model. The results show that based on the numerical simulator data, the initial productivity prediction accuracy of physics-guided XGBoost was verified as 94.67%, which was higher than the completely data-driven model. Moreover, the robustness of the novel adaptive physics-guided loss function for improving forecasting accuracy had been validated with sample sizes between 50 and 20000. By introducing the physics-guided sampling process and loss function, the applicability of the XGBoost algorithm has been extended to small sample conditions. Furthermore, the practical accuracy of physics-guided XGBoost was verified as 84.74% by 162 oil wells in the Penglai oilfield. Therefore, the application value of combining the physical knowledge with the XGBoost model has been confirmed in oilfield situations.

PO-03-226 MACHINE LEARNING IDENTIFIES FEATURES OF ARRHYTHMOGENIC QRS FRAGMENTATION IN HUMAN CARDIOMYOPATHY: IMPLICATIONS FOR IMPROVING AND AUTOMATING RISK STRATIFICATION

Slow, heterogeneous ventricular activation can provide the substrate for reentrant ventricular arrhythmias (VA), but its manifestation on the surface ECG as a risk stratifier is not well-defined beyond the presence of QRS duration (QRSd) >120ms or QRS fragmentation (fQRS). Our aim was to use machine learning (ML) to characterize the spatiotemporal features of QRS fragmentation that best predict VA in patients with cardiomyopathy (CM). Prospectively enrolled ischemic and non-ischemic CM patients with prophylactic defibrillators (n=95) underwent digital, high-resolution 12-lead ECG recordings for 5-minutes during intrinsic rhythm. Abnormal intra QRS peaks in the signal-averaged precordial leads (V1-V6) were identified using our previously validated time-domain algorithm. For each spatial location (V1-V6), QRS peak characteristics including width, amplitude and relative temporal location within the QRS were transformed into 4-bin histograms. Random forest models of the QRS peak counts and characteristics in each lead were trained on groups of 76 patients and tested on groups of 19 patients with stratified repeated random subsampling cross-validation to determine the features that predicted VA outcome. QRS peak counts and characteristics were additionally modelled with QRSd and other relevant clinical characteristics. CM patients (age 62±11 yrs, male 85%, LVEF 27±7%, ischemic etiology 62%) were followed for 24 (15-43) months and 22% had VA (+VA). QRSd was longer in +VA than -VA (135±25 vs. 114±26ms, p=0.001), but age, LVEF, CM etiology and fQRS were similar. The individual ML models determined (ROCAUC) QRS peak count (0.87), width (0.88) and location (0.84) to be predictors of VA (Table). In these models, the features most predictive of VA were a greater peak count in V1, a greater # of peaks less <6.25% of the QRSd in V1 and a greater # of peaks in the final quarter of the QRS of V1 (Figure). ML models of peak amplitude, QRSd and fQRS were not predictive of VA. The addition of clinical variables to the QRS peak count, width and location models did not improve VA prediction. Arrhythmogenic ventricular substrate can be independently identified during intrinsic rhythm by QRS fragmentation in V1 characterized by either a greater total # of peaks, a greater # of very narrow peaks, or a greater # of peaks at the end of the QRS according to our ML model. These findings require external validation and may improve risk stratification in defibrillator-eligible CM patients.Tabled 1Table. Performance of Random Forest ML Model.ModelROC [95% CI]Prespecified SensitivitySpecificityPPVNPVF1-ScoreQRS duration0.57 [0.35-0.87]90%0.220.270.490.42Qualitative fQRS0.56 [0.33-0.73]90%0.120.240.350.38All clinical characteristics∗0.46 (0.22-0.70]90%0.220.260.930.41QRS peak characteristics- Total Count0.87 [0.65-1.00]90%0.640.501.000.64- Width0.88 [0.70-1.00]90%0.730.541.000.69- Amplitude0.66 [0.36-0.93]90%0.250.290.710.44- Temporal Location in QRS0.84 [0.66-1.00]90%0.630.471.000.62QRS peak + clinical characteristics*- Total Count0.79 [0.55-0.97]90%0.560.421.000.57- Width0.82 [0.58-0.97]90%0.630.451.000.61- Amplitude0.68 [0.42-0.93]90%0.360.320.960.48- Temporal Location in QRS0.76 [0.52-0.95]90%0.530.391.000.55Age, Sex, Cardiomyopathy, QRS duration, Heart Rate, LVEF, NYHA Class, Comorbidities and Cardiac Medications Open table in a new tab

Management options influence seasonal CO2 soil emissions in Mediterranean olive ecosystems

Field trials were conducted at traditional Mediterranean olive agro-ecosystems grown at two locations (Italy –IT, Greece –GR). Groves were managed for many years using sustainable (S, cover crops, compost application, mulching of pruning biomass) or conventional (C) practices (e.g., soil tillage, burning of pruning residuals). The IT grove was rainfed (rain) while the GR was irrigated (IRR). This study examined the seasonal variation of soil CO2 emission (Rs) to explore the effect of the management options (C, S) on Rs at both sites. The second aim was to test the hypothesis that the seasonal Rs is differentially modulated by soil temperature and moisture, namely that (i) soil moisture limits Rs when it is below the lower limit of the readily available water (RAWLLim) and (ii) soil temperature above a threshold (max_T) reduces Rs even if soil moisture is non limiting. On the whole-season basis, the mean Rs rate at the rainfed site was 2.17 ± 0.06 (SE) at Crain and 2.32 ± 0.06 μmol CO2 m−2 s–1 at Srain plot, while at the irrigated site Rs was about 3.64 ± 0.11 (CIRR) and 4.05 ± 0.15 μmol CO2 m−2 s–1 (SIRR). The seasonal oscillation of Rs was consistent across locations and partitionable in three periods according to DOY (Day of Year) interval: Phase I (DOY 20–103 –GR; 20–118 -IT), Phase II (DOY 141÷257, GR; 142–257, IT) and Phase III (DOY 291–357, GR; 286–350, -IT). Pooling all the Rs data across sites and managements, max_T was ∼ 20 °C discriminating a differential response of Rs when soil moisture was < or > RAWLLim. These differential modulations exerted by temperature and moisture were integrated into a conditional model developed with a repeated random subsampling cross-validation procedure to effectively (R2 = 0.84) predict Rs. This paper mechanistically describes the interaction of the environment (soil moisture and temperature) and the management options (S, C) under various moisture conditions on Rs and would support carbon flux accounting procedures (e.g., regulating ecosystem services) tailored to the estimation of sink/source capability of traditional olive agro-ecosystem within environmental-friendly agricultural domains.

Machine-Learning Classification of SAR Remotely-Sensed Sea-Surface Petroleum Signatures—Part 1: Training and Testing Cross Validation

Sea-surface petroleum pollution is observed as “oil slicks” (i.e., “oil spills” or “oil seeps”) and can be confused with “look-alike slicks” (i.e., environmental phenomena, such as low-wind speed, upwelling conditions, chlorophyll, etc.) in synthetic aperture radar (SAR) measurements, the most proficient satellite sensor to detect mineral oil on the sea surface. Even though machine learning (ML) has become widely used to classify remotely-sensed petroleum signatures, few papers have been published comparing various ML methods to distinguish spills from look-alikes. Our research fills this gap by comparing and evaluating six traditional techniques: simple (naive Bayes (NB), K-nearest neighbor (KNN), decision trees (DT)) and advanced (random forest (RF), support vector machine (SVM), artificial neural network (ANN)) applied to different combinations of satellite-retrieved attributes. 36 ML algorithms were used to discriminate “ocean-slick signatures” (spills versus look-alikes) with ten-times repeated random subsampling cross validation (70-30 train-test partition). Our results found that the best algorithm (ANN: 90%) was &gt;20% more effective than the least accurate one (DT: ~68%). Our empirical ML observations contribute to both scientific ocean remote-sensing research and to oil and gas industry activities, in that: (i) most techniques were superior when morphological information and Meteorological and Oceanographic (MetOc) parameters were included together, and less accurate when these variables were used separately; (ii) the algorithms with the better performance used more variables (without feature selection), while lower accuracy algorithms were those that used fewer variables (with feature selection); (iii) we created algorithms more effective than those of benchmark-past studies that used linear discriminant analysis (LDA: ~85%) on the same dataset; and (iv) accurate algorithms can assist in finding new offshore fossil fuel discoveries (i.e., misclassification reduction).

Machine Learning–Based Text Analysis to Predict Severely Injured Patients in Emergency Medical Dispatch: Model Development and Validation

BackgroundEarly recognition of severely injured patients in prehospital settings is of paramount importance for timely treatment and transportation of patients to further treatment facilities. The dispatching accuracy has seldom been addressed in previous studies.ObjectiveIn this study, we aimed to build a machine learning–based model through text mining of emergency calls for the automated identification of severely injured patients after a road accident.MethodsAudio recordings of road accidents in Taipei City, Taiwan, in 2018 were obtained and randomly sampled. Data on call transfers or non-Mandarin speeches were excluded. To predict cases of severe trauma identified on-site by emergency medical technicians, all included cases were evaluated by both humans (6 dispatchers) and a machine learning model, that is, a prehospital-activated major trauma (PAMT) model. The PAMT model was developed using term frequency–inverse document frequency, rule-based classification, and a Bernoulli naïve Bayes classifier. Repeated random subsampling cross-validation was applied to evaluate the robustness of the model. The prediction performance of dispatchers and the PAMT model, in severe cases, was compared. Performance was indicated by sensitivity, specificity, positive predictive value, negative predictive value, and accuracy.ResultsAlthough the mean sensitivity and negative predictive value obtained by the PAMT model were higher than those of dispatchers, they obtained higher mean specificity, positive predictive value, and accuracy. The mean accuracy of the PAMT model, from certainty level 0 (lowest certainty) to level 6 (highest certainty), was higher except for levels 5 and 6. The overall performances of the dispatchers and the PAMT model were similar; however, the PAMT model had higher accuracy in cases where the dispatchers were less certain of their judgments.ConclusionsA machine learning–based model, called the PAMT model, was developed to predict severe road accident trauma. The results of our study suggest that the accuracy of the PAMT model is not superior to that of the participating dispatchers; however, it may assist dispatchers when they lack confidence while making a judgment.

Development and validation of early prediction models for new-onset functional impairment at hospital discharge of ICU admission.

We aimed to develop and validate models for predicting new-onset functional impairment after intensive care unit (ICU) admission with predictors routinely collected within 2days of admission. In this multi-center retrospective cohort study of acute care hospitals in Japan, we identified adult patients who were admitted to the ICU with independent activities of daily living before hospitalization and survived for at least 2days from April 2014 to October 2020. The primary outcome was functional impairment defined as Barthel Index ≤ 60 at hospital discharge. In the internal validation dataset (April 2014 to March 2019), using routinely collected 94 candidate predictors within 2 days of ICU admission, we trained and tuned the six conventional and machine-learning models with repeated random sub-sampling cross-validation. We computed the variable importance of each predictor to the models. In the temporal validation dataset (April 2019 to October 2020), we measured the performance of these models. We identified 19,846 eligible patients. Functional impairment at discharge was developed in 33% of patients (n = 6488/19,846). In the temporal validation dataset, all six models showed good discrimination ability with areas under the curve above 0.86, and the differences among the six models were negligible. Variable importance revealed newly detected early predictors, including worsened neurologic conditions and catabolism biomarkers such as decreased serum albumin and increased blood urea nitrogen. We successfully developed early prediction models of new-onset functional impairment after ICU admission that achieved high performance using only data routinely collected within 2days of ICU admission.

Early recognition of a caller’s emotion in out-of-hospital cardiac arrest dispatching: An artificial intelligence approach

AimThis study aimed to develop an AI model for detecting a caller’s emotional state during out-of-hospital cardiac arrest calls by processing audio recordings of dispatch communications. MethodsAudio recordings of 337 out-of-hospital cardiac arrest calls from March-April 2011 were retrieved. The callers’ emotional state was classified based on the emotional content and cooperative scores. Mel-frequency cepstral coefficients extracted essential information from the voice signals. A support vector machine was utilised for the automatic judgement, and repeated random sub-sampling cross validation (RRS-CV) was applied to evaluate robustness. The results from the artificial intelligence classifier were compared with the consensus of expert reviewers. ResultsThe audio recordings were classified into five emotional content and cooperative score levels. The proposed model had an average positive predictive value of 72.97%, a negative predictive value of 93.47%, sensitivity of 38.76%, and specificity of 98.29%. If only the first 10 seconds of the recordings were considered, it had an average positive predictive value of 84.62%, a negative predictive value of 93.57%, sensitivity of 52.38%, and specificity of 98.64%. The artificial intelligence model’s performance maintained preferable results for emotionally stable cases. ConclusionArtificial intelligence models can possibly facilitate the judgement of callers’ emotional states during dispatch conversations. This model has the potential to be utilised in practice, by pre-screening emotionally stable callers, thus allowing dispatchers to focus on cases that are judged to be emotionally unstable. Further research and validation are required to improve the model's performance and make it suitable for the general population.

Near-infrared spectroscopy outperforms genomics for predicting sugarcane feedstock quality traits.

The main objectives of this study were to evaluate the prediction performance of genomic and near-infrared spectroscopy (NIR) data and whether the integration of genomic and NIR predictor variables can increase the prediction accuracy of two feedstock quality traits (fiber and sucrose content) in a sugarcane population (Saccharum spp.). The following three modeling strategies were compared: M1 (genome-based prediction), M2 (NIR-based prediction), and M3 (integration of genomics and NIR wavenumbers). Data were collected from a commercial population comprised of three hundred and eighty-five individuals, genotyped for single nucleotide polymorphisms and screened using NIR spectroscopy. We compared partial least squares (PLS) and BayesB regression methods to estimate marker and wavenumber effects. In order to assess model performance, we employed random sub-sampling cross-validation to calculate the mean Pearson correlation coefficient between observed and predicted values. Our results showed that models fitted using BayesB were more predictive than PLS models. We found that NIR (M2) provided the highest prediction accuracy, whereas genomics (M1) presented the lowest predictive ability, regardless of the measured traits and regression methods used. The integration of predictors derived from NIR spectroscopy and genomics into a single model (M3) did not significantly improve the prediction accuracy for the two traits evaluated. These findings suggest that NIR-based prediction can be an effective strategy for predicting the genetic merit of sugarcane clones.

Comparison of artificial neural networks learning methods to evaluate supply chain performance

Abstract: The supply chain performance evaluation is a critical activity to continuously improve operations. Literature presents several performance evaluation systems based on multi-criteria methods and artificial intelligence. Among them, the systems based on artificial neural networks (ANN) excel due to their capacity of modeling non-linear relationships between metrics and allowing adaptations to a specific environment by means of historical performance data. These systems’ accuracy depend directly on the adopted training algorithm, and no studies have been found that assess the efficiency of these algorithms when applied to supply chain performance evaluation. In this context, the present study evaluates four ANNs learning methods in order to investigate which one is the most adequate to deal with supply chain evaluation. The algorithms tested were Gradient Descendent Momentum, Levenberg-Marquardt, Quasi-Newton and Scale Conjugate Gradient. The performance metrics were extracted from SCOR®, which is a reference model used worldwide. The random sub-sampling cross-validation method was adopted to find the most adequate topological configuration for each model. A set of 80 topologies was implemented using MATLAB®. The prediction accuracy evaluation was based on the mean square error. For the four level 1 metrics considered, the Levenberg-Marquardt algorithm provided the most precise results. The results of correlation analysis and hypothesis tests reinforce the accuracy of the proposed models. Furthermore, the proposed computational models reached a prediction accuracy higher than previous approaches.

Patient-specific PTV margins for liver stereotactic body radiation therapy determined using support vector classification with an early warning system for margin adaptation.

An adaptive planning target volume (PTV) margin strategy incorporating a volumetric tracking error assessment after each fraction is proposed for robotic stereotactic body radiation therapy (SBRT) liver treatments. A supervised machine learning algorithm employing retrospective data, which emulates a dry-run session prior to planning, is used to investigate if motion tracking errors are <2mm, and consequently, planning target volume (PTV) margins can be reduced. A fraction of data collected during the beginning of a treatment course emulates a dry-run session (mock) before planning. Twenty features are calculated using mock data and used for support vector classification (SVC). A treatment course is labeled as Class 1 if the maximum root-mean-square radial tracking error for all remaining fractions is below 2mm, or Class 2 otherwise. We evaluate the classification using fivefold cross-validation, leave-one-out cross-validation, 500 repeated random subsampling cross-validation, and the receiver operating characteristic (ROC) metric. The classification is independently cross-validated on a cohort of 48 treatment plans for other anatomical sites. A per fraction assessment of volumetric tracking errors is performed for the standard 5mm PTV margin (PTVstd ) for courses predicted as Class 2; or for a margin reduced by 2mm (PTVstd-2mm ) for those predicted as Class 1. We perturb the gross tumor volume (GTV) by the tracking errors for each x-ray image acquisition and calculate the fractional GTV voxel occupancy probability (Pi ) inside the PTV for each treatment fraction i. For treatment courses classified as Class 1, an early warning system flags treatment courses having any Pi <0.99, and the subsequent treatments are proposed to be replanned using PTVstd . The classification accuracies are 0.84±0.06 using fivefold cross-validation, and 0.77 when validated using an independent testing set (other anatomical sites). Eighty percent of treatment courses are correctly classified using leave-one-out cross-validation. The sensitivity, precision, specificity, F1 score, and accuracy are 0.81±0.09, 0.85±0.08, 0.80±0.11, 0.83±0.06, and 0.80±0.07, respectively, using 500 repeated random subsampling cross-validation. The area under the curve for the ROC metric is 0.87±0.05. The four most important features for classification are related to standard deviations of motion tracking errors, the linearity between the target location and external LED marker positions, and marker radial motion amplitudes. Eleven of 64 cases predicted to be of Class 1 have 0.96<Pi <0.99 for each treatment fraction, and require replanning using PTVstd . In comparison, the PTVstd always covers the perturbed GTVs with Pi >0.99 for all patients. Support vector classification is proposed for the classification of different motion tracking errors for patient courses based on a mock session before planning for SBRT liver treatments. It is feasible to implement patient-specific PTV margins in the clinic, assisted with an early warning system to flag treatment courses that require replanning using larger PTV margins in an adaptive treatment strategy.

Diagnostic Value of Combinatorial Markers in Colorectal Carcinoma.

Objectives: Blood-based tests have been shown to be an effective strategy for colorectal cancer (CRC) detection in screening programs. This study was aimed to test the performance of 20 blood markers including tumor antigens, inflammatory markers, and apolipoproteins as well as their combinations.Methods: In total 203 healthy volunteers and 102 patients with CRC were enrolled into the study. Differences between healthy and cancer subjects were evaluated using Wilcoxon rank-sum test. Several multivariate classification algorithms were employed using information about different combinations of biomarkers altered in CRC patients as well as age and gender of the subjects; random sub-sampling cross-validation was done to overcome overfitting problem. Diagnostic performance of single biomarkers and multivariate classification models was evaluated by receiver operating characteristic (ROC) analysis.Results: Of 20 biomarkers, 16 were significantly different between the groups (p-value ≤ 0.001); ApoA1, ApoA2 and ApoA4 levels were decreased, whereas levels of tumor antigens (e.g. carcinoembriogenic antigen) and inflammatory markers (e.g., C-reactive protein) were increased in CRC patients vs. healthy subjects. Combinatorial markers including information about all 16 significant analytes, age and gender of patients, demonstrated better performance over single biomarkers with average accuracy on test datasets ≥95% and area under ROC curve (AUROC) ≥98%.Conclusions: Combinatorial approach was shown to be a valid strategy to improve performance of blood-based CRC diagnostics. Further evaluation of the proposed models in screening programs will be performed to gain a better understanding of their diagnostic value.

An adaptive network-based fuzzy inference system to supply chain performance evaluation based on SCOR® metrics

Evaluating the performance of supply chains (SC) is a critical activity to enhance the outcomes of operations along the SC tiers. In order to support this evaluation process, several studies have proposed the application of artificial intelligence techniques combined with the performance metrics suggested by the SCOR® model (Supply Chain Operations Reference). However these propositions present some limitations. While the systems based on Mamdani fuzzy inference do not allow adaptation to the environment of use based on historical performance data, the systems based on artificial neural networks are not adequate to deal with imprecise data and qualitative metrics. In order to overcome these limitations, this paper presents a new approach to support SC performance evaluation based on the combination between the SCOR® metrics with an adaptive network-based fuzzy inference systems (ANFIS). In total, 56 candidate topologies were implemented and assessed using MATLAB. The random subsampling cross-validation method was applied to select the most appropriate topological parameters for each ANFIS model. The mean square error between the target values and the values estimated by each ANFIS model demonstrate its greater accuracy of prediction. In addition, results of the hypothesis tests based on paired samples indicate that the proposed ANFIS models are adequate to support SC performance evaluation. The proposed system can help managers to develop improvement actions plans based on the outcomes of the evaluation process. When compared to previous approaches, it presents advantages such as greater accuracy of prediction, learning ability based on historical data, suitability to support decision making under uncertainty, better interpretability of results, among others.

Juvenile fish habitat across the inner Danish waters: Habitat association models and habitat growth models for European plaice, flounder and common sole informed by a targeted survey

The management of coastal fisheries and coastal marine areas both require information on the use and importance of coastal marine habitats for marine species. This information is certainly pertinent for exploited species dependent on these coastal habitats for specific life-history stages. This study uses two different measures of habitat suitability to build Habitat Association Models (HAMs) and Habitat Growth Models (HGMs) for juveniles of three commercially and recreationally important fish species (Pleuronectes platessa, Platichthys flesus, and Solea solea) of the Inner Danish Waters (IDW). A large scale, one-off juvenile survey was used to collect juvenile fish and directly observed environmental data. The survey informed Generalised Linear (Mixed) Models (GLMs and GLMMs), which were fit and run through repeated random sub-sampling cross-validation before being utilised to build interpolative maps of both prediction and model uncertainty. The limitations of this approach are discussed in reference to the future works required to complement and improve such static habitat suitability maps. These results can be used to inform marine spatial planning and future work into the role of juvenile habitats in supporting fishery species.

Smudge Noise for Quality Estimation of Fingerprints and its Validation

Automated biometric identification systems are inherently challenged to optimize false (non-)match rates. This can be addressed either by directly improving comparison subsystems, or indirectly by allowing only “good quality” biometric queries to be compared. We are interested in the latter, where the challenge lies in relating the “good quality” of a query to its utility with respect to a comparison subsystem. First, we propose a new general robust biometric quality validation scheme (RBQ VS) that, mimicking the use-case, robustly quantifies comparison improvement obtained by employing a specific quality estimator. For this purpose, we robustify an existing validation scheme by repeated random subsampling cross-validation. Second, specifically for the task of fingerprint comparison, we propose a novel biometric feature for quality estimation. Since comparison subsystems based on fingerprint minutiae, which are ridge endings and bifurcations, appear to miss minutiae or detect spurious minutiae, especially in the presence of smudge noise, we propose an algorithm aiming at measuring corruption by smudge. To this end, we employ a recently developed three parts image-decomposition and link our new smudge noise quality estimator (SNoQE) to the structure of the texture part found. At last, using the FVC databases and an NIST database, we compare the SNoQE with the popular NFIQ 2.0 estimator, and its predecessor. Experimental results show that the single-feature SNoQE can compete with the multi-feature NFIQ 2.0 and, in fact, adds new information not sufficiently reproduced by the NFIQ 2.0. Indeed, a simple combination of SNoQE and NFIQ 2.0 tends to outperform on all databases included in the comparison study. An implementation of the RBQ VS and the SNoQE can be found online.

Predicting supply chain performance based on SCOR® metrics and multilayer perceptron neural networks

A supply chain performance prediction system aims to estimate lagging metrics based on leading metrics so as to predict performance based on causal relationships. Two studies in the literature propose a supply chain performance prediction system based on metrics suggested by the SCOR® (Supply Chain Operations Reference) model. However, a limitation of both systems is the difficulty of adjusting them to the environment of use, since their implementation and updating require manual parameterization of many fuzzy decision rules. To overcome this difficulty, this study proposes a performance prediction system also based on the SCOR® metrics but using artificial neural networks (ANN), which enables adaptation to a specific environment by means of historical performance data. Computational implementation of the ANN models was made using MATLAB. The method of random subsampling cross-validation was applied to select the network topologies. Results showed that the values of the correlation coefficient evidence that there is a high positive correlation between the expected and predicted performance values for the SCOR® level 1 metrics by all the ANN models. Statistical hypothesis tests showed that multilayer perceptron neural networks are adequate to support performance prediction of supply chains based on the SCOR® model. The proposed system promotes rational decision-making through a prospective diagnosis of the supply chain performance. By comparison between the predicted value and the target defined for each level 1 metric, managers can simulate whether improvement plans can lead to objectives; it can also help to identify areas that have performance problems and may need improvements.

Relatedness of Salmonella contamination frequency on chicken carcasses and parts when processed in the same establishment

Understanding how changes in the prevalence of Salmonella-positive chicken carcasses affects the prevalence of Salmonella-positive chicken parts samples across slaughter establishments is necessary to model the risk of foodborne illness. When developing new Salmonella performance standards for chicken parts—that were implemented following existing chicken carcass performance standards—FSIS made assumptions regarding the correlation between these two forms of chicken products marketed to consumers. These assumptions were necessary because data were not available at the time to measure this correlation. Using recent FSIS sampling data and covariate information concerning antimicrobials applied during chicken processing, regression analysis suggests a slight negative relationship between the prevalence of Salmonella-positive chicken parts in slaughter establishment and the use of peracetic acid on carcasses, as well as a large positive relationship between prevalence of Salmonella-positive chicken parts and carcasses. Nevertheless, using a repeated random sub-sampling cross-validation approach, the regression model has very limited predictive value. In lieu of a predictive model, estimation of the correlation between prevalence of Salmonella-positive chicken carcass and parts samples is useful for understanding the public health value of performance standards and other interventions applied to these chicken products. After adjusting the observed correlation of the sampling evidence to account for the underlying beta and binomial distribution errors inherent in observed results, the estimated correlation of prevalence of Salmonella-positive chicken carcass and parts is 0.54. This implies a moderate degree of relatedness but may still understate the true correlation because of limitations in the available data.

When the FAT goes wide: Right extended Frontal Aslant Tract volume predicts performance on working memory tasks in healthy humans.

The Frontal Aslant Tract (FAT) is a tract recently described as having implications on language function. The originally proposed anatomical FAT definition characterizes a connection between Broca’s territory and anterior supplementary and pre-supplementary motor areas in the Superior Frontal Gyrus (SFG). Here we propose an extended definition of the FAT (the exFAT) that propagates more anteriorly into the SFG. A sample of 834 subjects from the WU-Minn HCP 900 subjects data release (S900) was selected. The bilateral exFATs were reconstructed for the whole sample using an automated pipeline and thresholded adjusted tract volumes were calculated. A laterality test was performed on the whole sample. The frontal cortex has known implications on superior cognitive functions, so here we evaluate the implications of exFAT volume on performance in a language task and on a set of working memory tasks. Two sub-samples of 70 subjects each were drawn from the S900 sample by selecting the 35 top-performers and 35 bottom-performers for both language and working memory tasks. Additional laterality tests were performed on each subsample. We did not find the exFAT to be lateralized in any of the samples. We found statistically significant differences in left adjusted exFAT volume between top-performers and bottom-performers in the language task. We also found statistically significant differences in right adjusted exFAT volume between top-performers and bottom-performers for 2-back working memory tasks. To check for the predictive power of the exFAT volumes as correlates for performance, we ran a repeated random sub-sampling cross-validation procedure based on a Support Vector Machine (SVM) classifier that was capable of correctly classifying holdout subjects to their corresponding group (top-performer vs bottom-performer) with an average accuracy of 74.5% for language task performance based on left exFAT volume and an accuracy of 64.2% for Working Memory performance based on right exFAT volume.

Assessment of discriminant models in infrared imaging using constrained repeated random sampling – Cross validation

Infrared (IR) imaging is an emerging and powerful approach for studying the molecular composition of cells and tissues. It is a non-destructive and phenotypic technique which combines label-free molecular specific information from cells and tissues provided by IR with spatial resolution, offering great potential in biochemical and biomedical research and routine applications. The application of multivariate discriminant analysis using bilinear models such as Partial Least Squares-Discriminant Analysis (PLS-DA) to IR images requires to unfold the spatial directions in a two-way matrix, resulting in a loss of spatial information and structure. In this article, first we evidence that internal validation methods such as repeated k-fold cross-validation (CV) can be overly optimistic when the pixel size of the image is lower than the lateral spatial resolution. Secondly, we propose a new approach for the unbiased internal evaluation of the model performance named COnstrained Repeated Random Subsampling–Cross Validation (CORRS-CV). This method is based on the generation of q training and test sub-sets using a constrained random sampling of n training pixels without replacement and it circumvents overly optimistic effects due to oversampling, providing more accurate and robust images. The approach can be applied in IR microscopy for the development of discriminant models to analyse underlying biochemical differences associated to anatomical and histopathological features in cells and tissues.

Nondestructive, indirect assessment of the biomechanical properties of the rat intervertebral disc using contrast-enhanced μCT.

Mechanical characterization of the intervertebral disc involves labor-intensive and destructive experimental methodology. Contrast-enhanced micro-computed tomography is a nondestructive imaging modality for high-resolution visualization and glycosaminoglycan quantification of cartilaginous tissues. The purpose of this study was to determine whether anionic and cationic contrast-enhanced micro-computed tomography of the intervertebral disc can be used to indirectly assess disc mechanical properties in an ex vivo model of disc degeneration. L3/L4 motion segments were dissected from female Lewis rats. To deplete glycosaminoglycan, samples were treated with 0 U/ml (Control) or 5 U/ml papain. Contrast-enhanced micro-computed tomography was performed following incubation in 40% Hexabrix (anionic) or 30 mg I/ml CA4+ (cationic) for 24 h (n = 10/contrast agent/digestion group). Motion segments underwent cyclic mechanical testing to determine compressive and tensile modulus, stiffness, and hysteresis. Glycosaminoglycan content was determined using the dimethylmethylene blue assay. Correlations between glycosaminoglycan content, contrast-enhanced micro-computed tomography attenuation, and mechanical properties were assessed via the Pearson correlation. The predictive accuracy of attenuation on compressive properties was assessed via repeated random sub-sampling cross validation. Papain digestion produced significant decreases in glycosaminoglycan content and corresponding differences in attenuation and mechanical properties. Attenuation correlated significantly to glycosaminoglycan content and to all compressive mechanical properties using both Hexabrix and CA4+ . Predictive linear regression models demonstrated a predictive accuracy of attenuation on compressive modulus and stiffness of 79.8-86.0%. Contrast-enhanced micro-computed tomography was highly predictive of compressive mechanical properties in an ex vivo simulation of disc degeneration and may represent an effective modality for indirectly assessing disc compressive properties. © 2018 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 36:2030-2038, 2018.