SEARCH · PubMed Health

Results for “Model performance”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 109 records · Page 6Linked to original sources

PubMed1994

Young children's ability to understand a model as a spatial representation.

Children's ability to understand that a real environment can be represented in a symbolic form (i.e., by a model) is an important developmental achievement. Researchers have claimed that children who are just 3 years of age appreciate models as representations. This research was based on tasks that involved having young children use a model to locate a hiding place in an actual room. In this article, however, we point out the difficulties in interpreting previous model tasks, and we describe two studies that showed that young 3-year-olds could perform model tasks successfully when the hiding place they were looking for was a unique place in the model (and room). When the hiding place was unique, the children had to note only that place and they needed no further knowledge about the relationship between the model and the room. When the hiding place was one of two identical places, however, the children needed to take spatial relationships into account to distinguish the correct place, and young 3-year-olds were unable to do this. Four-year-olds were able to use spatial relationships to distinguish identical places when the model was aligned with the space it represented, but they had difficulty when the model was not aligned. Five-year-olds could use spatial relationships effectively between one model space and another whether or not the model was aligned.

Child↗

PubMed1985

Digital separation of primary and scatter components of chest radiographs.

This article describes a technique for digital separation of the primary and scatter components of a radiographic image. The method involves mathematical modeling of the process whereby an antiscatter grid reduces scatter patterns in film radiographs. Two superimposable radiographs (one taken with and the other without an intervening antiscatter grid) are applied to the model. Performance characteristics of the grid (primary and scatter transmittance factors) are also determined and used in the model. Radiographs of a humanoid chest phantom are processed. Scatter/primary separation appears to be accurate to within 15%. Film images that are quantitatively faithful to the calculated primary and scatter fields are included.

Analog-Digital Conversion↗

PubMed2025

AWGE-ESPCA: An edge sparse PCA model based on adaptive noise elimination regularization and weighted gene network for Hermetia illucens genomic data analysis.

Hermetia illucens is an important insect resource. Studies have shown that exploring the effects of Cu2+-stressed on the growth and development of the Hermetia illucens genome holds significant scientific importance. There are three major challenges in the current studies of Hermetia illucens genomic data analysis: firstly, the lack of available genomic data which limits researchers in Hermetia illucens genomic data analysis. Secondly, to the best of our knowledge, there are no Artificial Intelligence (AI) feature selection models designed specifically for Hermetia illucens genome. Unlike human genomic data, noise in Hermetia illucens data is a more serious problem. Third, how to choose those genes located in the pathway enrichment region. Existing models assume that each gene probe has the same priori weight. However, researchers usually pay more attention to gene probes which are in the pathway enrichment region. Based on the above challenges, we initially construct experiments and establish a new Cu2+-stressed Hermetia illucens growth genome dataset. Subsequently, we propose AWGE-ESPCA: an edge Sparse PCA model based on adaptive noise elimination regularization and weighted gene network. The AWGE-ESPCA model innovatively proposes an adaptive noise elimination regularization method, effectively addressing the noise challenge in Hermetia illucens genomic data. We also integrate the known gene-pathway quantitative information into the Sparse PCA(SPCA) framework as a priori knowledge, which allows the model to filter out the gene probes in pathway-rich regions as much as possible. Ultimately, this study conducts five independent experiments and compared four latest Sparse PCA models as well as representative supervised and unsupervised baseline models to validate the model performance. The experimental results demonstrate the superior pathway and gene selection capabilities of the AWGE-ESPCA model. Ablation experiments validate the role of the adaptive regularizer and network weighting module. To summarize, this paper presents an innovative unsupervised model for Hermetia illucens genome analysis, which can effectively help researchers identify potential biomarkers. In addition, we also provide a working AWGE - ESPCA model code in the address: https://github.com/yhyresearcher/AWGE_ESPCA.

Animals↗

PubMed2025

Predicting natural variation in the yeast phenotypic landscape with machine learning.

Most organismal traits result from the complex interplay of many genetic and environmental factors, making their prediction difficult. Here, we used machine learning (ML) models to explore phenotype predictions for 223 traits measured across 1011 genome-sequenced Saccharomyces cerevisiae strains isolated worldwide. We benchmarked a ML pipeline with multiple linear and non-linear models to predict phenotypes from genotypes and gene expression, and determined gradient boosting machines as the best-performing model. Gene function disruption scores and gene presence/absence emerged as best predictors, suggesting a considerable contribution of the accessory genome in controlling phenotypes. The prediction accuracy broadly varied among phenotypes, with stress resistance being easier to predict compared to growth across nutrients. ML identified relevant genomic features linked to phenotypes, including high-impact variants with established relationships to phenotypes, despite these being rare in the population. Near-perfect accuracies were achieved when other phenomics data mostly in similar conditions were used, suggesting that useful information can be conveyed across phenotypes. Overall, our study underscores the power of ML to interpret the functional outcome of genetic variants.

Genetic Variation↗

PubMed1993

Quality of laboratory performance in testing for human immunodeficiency virus type 1 antibody. Identification of variables associated with laboratory performance.

To identify factors that may affect the quality of laboratory performance of human immunodeficiency virus type 1 (HIV-1) antibody testing, the Centers for Disease Control and Prevention Model Performance Evaluation Program surveyed laboratories in 1989 that performed enzyme immunoassay (EIA) and Western blot tests for HIV-1 antibody. Panels of 10 HIV-1-antibody-positive and antibody-negative plasma samples, some of which were duplicates, were mailed to program-participating laboratories. Laboratories were also mailed survey questionnaires to ascertain their laboratory characteristics and testing practices. Using 1988 data, researchers previously found that the overall analytic performance of laboratories performing HIV-1 antibody testing was independently associated with the following: (1) requiring a minimum degree of testing personnel; (2) having written criteria for identifying unsatisfactory specimens; (3) requiring in-house training for testing personnel; (4) having tested more than 10,000 specimens; (5) being identified as an "other" laboratory type; (6) having more than 24 months of testing experience; (7) laboratory uses specific (Abbott) materials for EIA; and (8) testing specimens collected by family-planning clinics. To verify these findings, we performed multivariate analysis on 1989 performance data. For the 1989 EIA analytic sensitivity, significant positive (P < or = .05) associations were detected with having written criteria for identifying unsatisfactory specimens and with having tested more than 10,000 specimens. For the 1989 overall EIA analytic performance, a significant negative (P < or = .05) association was found with using specific (Abbott) EIA materials, and a significant positive (P < or = .05) association was found with having tested more than 10,000 specimens. For Western blot results, the only significant (P < or = .05) associations were for both analytic sensitivity and overall analytic performance and having tested more than 10,000 specimens.

Blotting, Western↗

PubMed2026

Genetic Susceptibility to Incisional Hernia Evaluation of Hernia Polygenic Risk Scores.

OBJECTIVES: Incisional hernia (IH) affects 13-30% of people after abdominal surgery, resulting in substantial morbidity and costs. While clinical risk factors have been studied extensively, genomic risk for IH is incompletely understood. We aimed to evaluate the impact of polygenic risk scores (PRS) on IH risk prediction. METHODS: We created and evaluated three PRS for abdominal hernia, ventral hernia and latent hernia susceptibility for prediction of IH in an institutional biobank. The primary outcome was defined as the diagnosis or repair of an IH based on ICD-9/10-CM/PCS and CPT codes. Clinical covariates included age, sex, body mass index (BMI), smoking status, index procedure type, and perioperative surgical site infection. A phenome-wide association study (PheWAS) was performed to assess clinical associations with increased PRS. We then tested the ability of the PRS to improve prediction for IH by modeling clinical covariates with and without PRS in patients who underwent abdominal surgery. Model performance was assessed using 10 iterations of 5-fold cross-validation to estimate Brier scores and area under the receiver operating characteristic curve (AUROC), which were compared using cross-model Bayesian analysis of variance. RESULTS: In 55,809 subjects, assessed PRS was significantly associated with incisional, umbilical, and ventral hernia on PheWAS, with 1.19 greater odds of developing IH per 1-SD increase in PRS (95% CI: 1.13-1.25, P < 0.001). Of 9,909 subjects who underwent qualifying abdominal surgery, 706 developed IH. In this cohort, the latent hernia susceptibility PRS was associated with a 16% increased hazard of developing IH per 1-SD increase (HR 1.16; 95% CI: 1.07-1.26; P < 0.001). Compared to a predictive model using clinical covariates (Brier score = 0.047, 95% CI: 0.046-0.048; AUROC = 0.660, 95% CI: 0.653-0.666), addition of the PRS showed similar Brier score and AUROC estimates (Brier score = 0.047, 95% CI: 0.046-0.048; AUROC: 0.667, 95% CI: 0.661-0.673) at five years. Cross-model Bayesian analysis demonstrated >99% probability of practical equivalence when trying to detect a difference of ≥ 0.02. CONCLUSION: All three PRS for hernia were independently associated with IH, suggesting that genomic factors contribute significantly to IH development. However, none of the three PRS meaningfully improved clinical IH risk prediction in patients who underwent abdominal surgery. This suggests that clinical comorbidities and surgical techniques may be equally as important as genomic architecture.

Bayesian analysis↗

PubMed1991

Adequacy of a systems structure in the modeling of training effects on performance.

A systems model of training effects on performance was applied to eight initially untrained subjects who were volunteers for an endurance training program for the purpose of verifying the statistical adequacy of the systems structure. In the model initially proposed by T. W. Calvert, E. W. Banister, M. V. Savage, and T. Bach (IEEE Trans. Syst. Man Cybern. 6: 94-102, 1976), the performance changes were related to the successive training loads by three first-order transfer functions. In the present study, the number of first-order components was statistically tested. A model including only one component, which had a positive effect on the performance, provided a significant fit with the performances in every subject. A second component significantly improved the fit in only two subjects. This further component, which had a negative effect on performance, was identified as fatigue. Nevertheless, a two-antagonistic component model is proposed to provide a good representation of the training responses. However, the low level of exercise demands and the inaccuracy of the fit could have impaired the evidencing of a fatiguing effect during the presently studied training protocol.

Adult↗

PubMed1981

The invariance of sentence performance structures across language modality.

Native users of American Sign Language were asked to manipulate sentences in four different ways: sign them at slow rate, parse them, make relatedness judgments of pairs of signs taken from each sentence, and recall the sentences. The data obtained from these four tasks (pause durations, parsing values, indices of relatedness and probe latencies) were used to construct hierarchical performance structures for each of the sentences. The resulting structures were highly similar across tasks; that is, performance structures are not task specific. The four measures at each sign boundary in each sentence were well predicted by a performance model, elaborated by Grosjean, Grosjean, and Lane for speech, that combines a parsing measure with a symmetry measure. Thus performance structures appear to be founded in the processing of language, be it visual or oral, and not in the properties of any particular communication modality.

Adult↗

PubMed2025

Examining Transcriptomic Markers Associated With Neutrophil Extracellular Traps to Predict Mortality Risk in Neonatal Sepsis.

BACKGROUND: Neonates are highly susceptible to sepsis, which is often accompanied by fatal coagulopathy. Anticoagulant therapies have not reduced sepsis-related mortality in clinical trials, possibly due to patient heterogeneity. Neutrophil extracellular traps (NETs) enhance coagulation by activating platelets, suggesting that NET-specific biomarkers may identify patients who may benefit from targeted anticoagulant treatment. This study evaluated the association between NET gene expression and adverse outcomes in neonatal sepsis. METHODS: We analyzed whole blood transcriptomes from 123 neonates with sepsis and developed a predictive model, the NET score, based on NET-related gene expression. Model performance was assessed in two independent validation sets. Mediation and correlation analyses explored the relationship between the NET score and a coagulation score. Temporal transcriptomic data from septic shock cases further tested this interaction. RESULTS: The NET score achieved AUCs of 88.7% and 85.4% in validation Sets 1 and 2, respectively, indicating strong predictive performance. Mediation and temporal analyses supported a sequential relationship between NETosis and coagulation in sepsis. Age-specificity of the model was confirmed using pediatric (n = 163) and adult (n = 86) sepsis transcriptomic datasets. Neonates with disseminated intravascular coagulation exhibited a trend toward elevated NET scores. CONCLUSIONS: Our findings support a novel risk stratification approach using the NET score to identify neonates at increased risk for sepsis-associated coagulopathy and poor outcomes, potentially guiding targeted therapeutic strategies.

neonatal sepsis↗

PubMed1992

Monte Carlo modelling of the performance of a rotating slit-collimator for improved planar gamma-camera imaging.

Planar imaging with a gamma camera is currently limited by the performance of the collimator. Spatial resolution and sensitivity trade off against each other; it is not possible with conventional parallel-hole collimation to have high geometric sensitivity and at the same time excellent spatial resolution unless field-of-view is sacrificed by using fan- or cone-beam collimators. We propose a rotating slit-collimator which collects one-dimensional projections from which the planar image may be reconstructed by the theory of computed tomography. The performance of such a collimator is modelled by Monte Carlo methods and images are reconstructed by a convolution and backprojection technique. The performance is compared with that of a conventional parallel-hole collimator and it is shown that higher spatial resolution with increased sensitivity is possible with the slit-collimator. For a point source a spatial resolution of some 6 mm at a distance of 100 mm from the collimator with a x7 sensitivity compared with a parallel-hole collimator was achieved. Applications to bone scintigraphy are modelled and an improved performance in hot-spot imaging is demonstrated. The expected performance in cold-spot imaging is analytically investigated. The slit-collimator is not expected to improve cold-spot imaging. Practical design considerations are discussed.

Equipment Design↗

PubMed1992

Assessment of a model for overall left ventricular three-dimensional motion from MRI data.

In this paper we present a complete methodology to evaluate a model for overall three-dimensional (3D) motion of the human left ventricle (LV) from MRI data. The left ventricular motion is approximated by a linear model associated with an affine transformation to determine parameters for non-rigid motion of the LV. The proposed method has been applied to a normal patient and to a patient with cardiac disease. Results obtained show that the linear model provides a fairly good approximation of normal left ventricular motion, whereas serious cardiac disease produces abnormal motion, yielding altered model performances.

Heart Diseases↗

PubMed1995

Statistical response models for ozone exposure: their generality when applied to human spirometric and animal permeability functions of the lung.

Exposure of humans or animals to ozone (O3) alters spirometric and permeability functions of the lung. While these responses show clear concentration (C) dependency, the interactive role of exposure duration (T) has not been well defined. Ozone-induced alterations in forced expiratory volume in 1 s (FEV1) obtained from human studies and in levels of bronchoalveolar lavage fluid protein (BALP) obtained from studies of rats and guinea pigs were used to compare the utility of several proposed response models as functions of C and T. A large human-study database compiled for T = 2 h and a wide-ranging C and T study on animals were used to contrast each model. The models examined included the quadratic, logistic, log regression, and exponential models. This work suggests that models used for risk assessment should incorporate both T and C. Our results suggest that modified forms of many of these models perform well with both human and animal responses and can be additionally modified to include ventilation rate. As a simple biological model, the exponential model showed advantages. The absolute concentration rates-of-change in the exponential function of integrated physiological changes like BALP and FEV1 were equal for low O3 exposure.

Animals↗

PubMed1994

The identification of Class III malocclusions by discriminant analysis.

Lateral cephalometric radiographs of 210 control and 285 Class III subjects were traced, digitized, and 43 calculated variables submitted to a stepwise discriminant analysis. A 10-factor model was generated, giving 95.2 per cent correct classification of the control children and 95.1 per cent accurate identification of the Class III group. Ten control and 14 Class III children were categorized incorrectly. To test the validity of the analysis, radiographs of these 24 individuals were examined in detail. In all cases, a satisfactory reason for the misgrouping was identified. This investigation underlined the importance of rigorous standards of case selection when compiling the groups. The robustness of the 10-factor model was also examined. Subjects were arbitrarily split into two groups, the odd- and the even-numbered cases, and the discriminant analysis repeated on each. Both new models contained the same 10 variables, but with slightly different values for their accompanying coefficients. The cases erroneously identified by the whole group analysis were again misclassified, together with a few additional cases. Each new model performed equally well on the data from the opposing group as on that from which it had been derived. Thus, the model generated in this study was both valid and acceptably robust. It would therefore appear that discriminant analysis may be a viable tool in the identification and classification of groups of individuals.

Adolescent↗

PubMed1995

Predicting future functional status for seriously ill hospitalized adults. The SUPPORT prognostic model.

OBJECTIVE: To develop a model estimating the probability of an adult patient having severe functional limitations 2 months after being hospitalized with one of nine serious illnesses. DESIGN: Prospective cohort study. SETTING: Five teaching hospitals in the United States. PARTICIPANTS: 1746 patients (model development) who survived 2 months and completed an interview, selected from 4301 patients in the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT); independent validation sample of 2478 patients. MEASUREMENTS AND MAIN OUTCOMES: Patient function 2 months after admission categorized as absence or presence of severe functional limitations (defined as Sickness Impact Profile scores > or = 30 or as activities of daily living scores > or = 4 [levels that require near-constant personal assistance]). A logistic regression model was constructed to predict severe functional limitation. RESULTS: One third (n = 590) of patients who were interviewed at 2 months had severe functional limitations. Changes in functional status were common: Of those with no baseline dependencies (not dependent on personal assistance), 21% were severely limited at 2 months; of those with 4 or more baseline limitations, 30% had improved. The patient's ability to do activities of daily living was the most important predictor of functional status. Physiologic abnormalities, diagnosis, days in hospital, age, quality of life, and previous exercise capacity also contributed substantially. Model performance, assessed using receiver-operating characteristic curves, was 0.79 for the development sample and 0.75 for the validation sample. The model was well calibrated for the entire risk range. CONCLUSIONS: Functional outcome varied substantially after hospitalization for a serious illness. A small amount of readily available clinical information can estimate the probability of severe functional limitations.

Activities of Daily Living↗

PubMed1987

Collicular ensemble coding of saccades based on vector summation.

The superior colliculus in the monkey contains a topographically organized representation of the target in its upper layers and saccade-related activity in its deeper layers. Since collicular movement fields are quite large, a considerable region of the colliculus is active whenever a saccade is made. We have modelled the collicular role in saccade generation based on the idea, proposed earlier in the literature, that each movement cell causes a movement tendency in the direction of the external world point which it represents in the collicular map. The model is organized as follows: An anisotropic logarithmic mapping transforms retinal coordinates into collicular coordinates. A two-dimensional Gaussian function describes the spatial extent of the movement-related activity in the deeper layers. An efferent mapping function specifies how the direction and the size of the movement contribution of each colliculus neuron depends on its location and its firing rate. The total saccade is the vector sum of the individual cell contributions. This very simple model (seven fixed parameters) has been used to simulate metrical properties of saccades: in response to visual targets; in response to electrical stimulation in one colliculus, and after a colliculus lesion. Model performance appears to be remarkably realistic but cannot account for some border effects and responses to double stimulation. Suggestions on how the model can be improved and extended will be presented.

Brain Mapping↗

PubMed1975

Cost-performance analysis of cataloging and card production in a medical center library.

The unit cost of cataloging current English language monographs was studies and compared with the cost of purchasing catalog cards from a commercial source. Two hypotheses were proposed: (1) in-library costs for cataloging and card production are higher than those for the purchased-card method; (2) throughput time is faster for the in-library method. In addition, the data can be used to develop an analytical cost-performance model for administrative purposes. The data presented support the hypotheses. The model developed provides a mechanism for arriving at a cost for different levels of service and can be used to measure the performance of other alternative methods of cataloging. Implications for the use of CATLINE are discussed and suggestions for further studies are described.

Cataloging↗

PubMed2025

Opportunities for machine learning to predict cross-neutralization in FMDV serotype O.

Accurately estimating cross-neutralization between serotype O foot-and-mouth disease viruses (FMDVs) is critical for guiding vaccine selection and disease management. In this study, we developed a machine learning approach to estimate r1 values-an established measure of antigenic similarity-using VP1 sequence data and published virus neutralization titer (VNT) results. Our dataset comprised 108 serum-virus pairs representing 73 distinct FMDV strains. We applied Boruta feature selection and random forest classifiers, optimizing model performance through tenfold cross-validation and sub-sampling to address class imbalance. Predictors included pairwise amino acid distances, site-specific polymorphisms, and differences in potential N-glycosylation sites. Using a 0.3 r1 threshold to define cross-neutralization, the final model achieved high accuracy (0.96), sensitivity (0.93), and specificity (0.96) in training, and performed robustly on independent test sets - accuracy was 0.75 (95% CI 0.60 and 0.90), F1 score 0.86% and PPV 0.77. Importantly, key VP1 residues-positions 48, 100, 135, 150, and 151-emerged as strong predictors of antigenic relationships. Our results demonstrate the utility of integrating routinely generated genomic data with machine learning to inform vaccine candidate selection and anticipate immune interactions among circulating FMDV strains. This approach offers a practical tool for accelerating vaccine decision-making and can be adapted to other FMDV serotypes. The latest version of the r1 predictive model is available for access via a Shiny dashboard (https://dmakau.shinyapps.io/PredImmune-FMD/).

Foot-and-Mouth Disease Virus↗

PubMed2026

Benchmarking large language models for extracting biobank-derived insights into health and disease.

Biobank-scale datasets such as the UK Biobank have become foundational resources for advancing biomedical discovery. Yet the complexity and heterogeneity of these resources, spanning genomics, imaging, clinical records, and metadata, pose substantial barriers to access and interpretation. Large Language Models (LLMs) offer a promising avenue for making such datasets more navigable through natural language interfaces. However, the extent to which current general-purpose LLMs can retrieve and synthesize biobank-specific insights has not yet been systematically evaluated. In this study, we present a reproducible, multi-metric evaluation framework to benchmark the capabilities of leading LLMs. We evaluated six leading large language models: Gemini 3 Pro, Claude Opus 4.5, Claude Sonnet 4.5, GPT-5.2, Mistral Large 2, and DeepSeek V3, on four benchmark tasks designed to assess biobank-related knowledge retrieval. We evaluate model performance across six dimensions (semantic accuracy, factual correctness, domain knowledge, reasoning quality, response depth, and biobank specificity) and assessed output consistency using curated UK Biobank references and a robust random baseline. All models outperformed the baseline by 2× to 3× , with strong statistical separation (p < 0.001), confirming meaningful biobank-specific knowledge retrieval. Gemini 3 Pro achieved the highest overall accuracy across tasks such as keyword synthesis, institution recognition, and topic inference, while Claude Sonnet 4.5 demonstrated the most uniform performance across evaluation dimensions. Our benchmark provides a rigorous framework for evaluating LLMs in biomedical settings. Using the UK Biobank as a real-world testbed, we highlight both the capabilities and limitations of current models, measuring their capacity to recall structured biomedical knowledge consistent with authoritative biobank metadata.

Large Language Models↗