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 181 records · Page 10Linked to original sources

PubMed1997

Lateralized spatial strategies in oscillating drawing movements.

Kinematic characteristics and lateral differences between two upper extremities were investigated in a unimanual graphic task involving fast and precise oscillating movements on the vertical plane. The spatial locations of sequential reversal points were used to calculate the pairs of angles, relative to the horizontal axis. The point biserial coefficient of correlation was used to analyze the difference between big and large angles and their sequence in each pair. Three main groups (A, B, and C) of performance models were distinguished in 132 tests by 33 strongly right-handed male subjects. Group A showed strong variation in vertical movement, Group B covariation in vertical and horizontal vectors, while Group C reflected independent variation of both vertical and horizontal directions. It is suggested that the movement strategies might reflect three different models of motor control involving coupling of an oscillator controlling pools of motoneurons which regulates horizontal movements with an oscillator controlling vertical movement (Groups A + B) or with nonoscillating control signal (Group B). It is argued that Group A represents the simplest strategy and only performance Type A met by the left hand.

Adult↗

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↗

PubMed1998

Time series forecasts of ambulance run volume.

To test the hypothesis that time series analysis can provide accurate predictions of future ambulance service run volume, a prospective stochastic time series modeling study was conducted at a community-based regional ambulance service. For all requests for ambulance transport during two sequential years, the time and date, total run time, and acuity code of the run were recorded in a computer database. Time series variables were formed for ambulance service runs per hour, total run time, and acuity. Prediction models were developed from one complete year's data (1994) and included four model types: raw observations, moving average, means with moving average smoothing, and autoregressive integrated moving average. Forecasts from each model were tested against observations from the first 24 weeks of the subsequent year (1995). Each model's adequacy was tested on residuals by autocorrelation functions, integrated periodograms, linear regression, and differences among the variances. A total of 68,433 patients were seen in 1994 and 32,783 in the first 24 weeks of 1995. Large periodic variations in run volume with time of day were found (P < .001). A model based on arithmetic means of each hour of the week with 3-point moving average smoothing yielded the most accurate forecasts and explained 54.3% of the variation observed in the 1995 test series (P < .001). Time series analysis can provide powerful, accurate short-range forecasts of future ambulance service run volume. Simpler, less expensive models performed best in this study.

Ambulances↗

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↗

PubMed1990

Effect of visuomotor rehearsal with videotaped modeling on racquetball performance of beginning players.

This investigation was designed to assess the effect of visuomotor behavior rehearsal (VMBR) with videotaped modeling on racquetball performance of beginning players. 24 male students in beginning racquetball class were randomly assigned to either VMBR with videotape modeling or relaxation and imagery (no modeling) condition for a 2-wk. training period. All subjects were pre- and posttested on forehand and backhand racquetball skill tests. Analysis indicated an effect for forehand shooting, with those given visuomotor behavior rehearsal with videotaped model exhibiting better performance than the relaxation and imagery group, but no effect for backhand shooting.

Adolescent↗

PubMedDate not supplied

Nonparametric block-structured modeling of lung tissue strip mechanics.

Very large amplitude pseudorandom uniaxial perturbations containing frequencies between 0.125 and 12.5 Hz were applied to five dog lung tissue strips. Three different nonlinear block-structured models in nonparametric form were fit to the data. These models consisted of (1) a static nonlinear block followed by a dynamic linear block (Hammerstein model); (2) the same blocks in reverse order (Wiener model); and (3) the blocks in parallel (parallel model). Both the Hammerstein and Wiener models performed well for a given input perturbation, each accounting for greater than 99% of the measured stress signal variance. However, the Wiener and parallel model parameters showed some dependence on the strain amplitude and the mean stress. In contrast, a single Hammerstein model accounted for the data at all strain amplitudes and operating stresses. A Hammerstein model featuring a fifth-order polynomial static nonlinearity and a linear impulse response function of 1 s duration accounted for the most output variance (99.84%+/-0.13%, mean+/-standard deviations for perturbations of 50% strain at 1.5 kPa stress). The static nonlinear behavior of the Hammerstein model also matched the quasistatic stress-strain behavior obtained at the same strain amplitude and operating stress. These results show that the static nonlinear behavior of the dog lung tissue strip is separable from its linear dynamic behavior.

Animals↗

PubMed1997

Estimation of the DNA sequence discriminatory ability of hairpin-linked lexitropsins.

Three- and four-ring polyamides containing N-methylimidazole and N-methylpyrrole, and their hairpin-linked derivatives, bind side-by-side in the minor groove of DNA in a sequence-specific manner. The sequences recognized by side-by-side molecules are dependent on the pairings of the polyamide rings to the bases. In this study we report a mathematical model for estimating the free energies of binding for gamma-aminobutyric acid-linked polyamides to 5- and 6-bp DNA sequences. The model parameters are calibrated by a least-squares fit to 35 experimental binding constants. The model performs well in cross-validation experiments and the parameters are consistent with previously proposed empirical rules of polyamide-DNA binding. We apply the model to the design of targeted polyamides, evaluating the ability of the proposed polyamides to bind to a DNA sequence of interest while minimizing binding to the remaining DNA sequences.

Antineoplastic Agents↗

PubMed1997

Mathematical analysis of motion-opponent mechanisms used in the determination of heading and depth.

A mathematical analysis is presented of a model that uses motion-opponent operators similar to neurons found in the primate middle temporal visual area, to determine observer heading and depth from optical flow information. The response of these operators to depth changes in the form of a slanted plane or a step edge is analyzed, and the outputs of odd-symmetric operators are compared with that of circularly symmetric operators. The analysis shows sources of error from these operators in determining heading and depth and suggests how some of these errors can be mitigated. Simulations are presented that show that the model performs well for a variety of situations.

Animals↗

PubMed1995

Time series forecasts of poison center call volume.

We tested the hypothesis that time series analysis can provide accurate predictions of future poison center telephone call volume by a prospective stochastic time series modeling of calls to a university-based regional poison center. All callers evaluated and managed during two sequential years had the time and date of the call recorded in a computer database. Time series variables were formed for poison center calls per hour. Prediction models were developed from the 1992 data and included four types: raw observations, moving average, means with moving average smoothing, and autoregressive integrated moving average. Forecasts from each model were tested against observations from the first 26 weeks of 1993. Each model's adequacy was tested on residuals by autocorrelation functions, integrated periodograms, linear regression, and differences among the variances. A total of 44,584 calls were received in 1992 and 24,781 in the first half of 1993. Large periodic variations in call volume with time of day were found (p < 0.001). The model based on arithmetic means of each hour of the week with three-point moving average smoothing yielded the most accurate forecasts and explained 58.5% of the variation observed in the 1993 test series (p < 0.001). Time series analysis can provide powerful, accurate short range forecasts of future poison center telephone call volume. Simpler, less expensive models performed best in this study.

Computers↗

PubMed2020

Exploration of predictive and prognostic alternative splicing signatures in lung adenocarcinoma using machine learning methods.

BACKGROUND: Alternative splicing (AS) plays critical roles in generating protein diversity and complexity. Dysregulation of AS underlies the initiation and progression of tumors. Machine learning approaches have emerged as efficient tools to identify promising biomarkers. It is meaningful to explore pivotal AS events (ASEs) to deepen understanding and improve prognostic assessments of lung adenocarcinoma (LUAD) via machine learning algorithms. METHOD: RNA sequencing data and AS data were extracted from The Cancer Genome Atlas (TCGA) database and TCGA SpliceSeq database. Using several machine learning methods, we identified 24 pairs of LUAD-related ASEs implicated in splicing switches and a random forest-based classifiers for identifying lymph node metastasis (LNM) consisting of 12 ASEs. Furthermore, we identified key prognosis-related ASEs and established a 16-ASE-based prognostic model to predict overall survival for LUAD patients using Cox regression model, random survival forest analysis, and forward selection model. Bioinformatics analyses were also applied to identify underlying mechanisms and associated upstream splicing factors (SFs). RESULTS: Each pair of ASEs was spliced from the same parent gene, and exhibited perfect inverse intrapair correlation (correlation coefficient = - 1). The 12-ASE-based classifier showed robust ability to evaluate LNM status of LUAD patients with the area under the receiver operating characteristic (ROC) curve (AUC) more than 0.7 in fivefold cross-validation. The prognostic model performed well at 1, 3, 5, and 10 years in both the training cohort and internal test cohort. Univariate and multivariate Cox regression indicated the prognostic model could be used as an independent prognostic factor for patients with LUAD. Further analysis revealed correlations between the prognostic model and American Joint Committee on Cancer stage, T stage, N stage, and living status. The splicing network constructed of survival-related SFs and ASEs depicts regulatory relationships between them. CONCLUSION: In summary, our study provides insight into LUAD researches and managements based on these AS biomarkers.

Adenocarcinoma of Lung↗

PubMed1995

Afferent synaptic drive of rat medial nucleus tractus solitarius neurons: dynamic simulation of graded vesicular mobilization, release, and non-NMDA receptor kinetics.

1. We have developed a comprehensive mathematical model of an afferent synaptic connection to the soma of a medial nucleus tractus solitarius (mNTS) neuron. Model development is based on numerical fits to quantitative data recorded in our laboratory. This work is part of a continuing collaborative effort aimed at identifying and characterizing the mechanisms responsible for the non-linear integrative properties of this first synapse in the baroreceptor reflex. 2. The complete model consists of three major parts: 1) a Hodgkin-Huxley (HH)-type membrane model of the prejunctional sensory terminal bouton; 2) a multistage model describing vesicular storage, adenosine 3',5'-cyclic monophosphate (cAMP)- and Ca(2+)-dependent mobilization, release and recycling; and 3) a HH-type membrane model of the postjunctional mNTS cell that includes descriptions for a desensitizing non-N-methyl-D-aspartate (NMDA) ionic current that is responsible for the fast excitatory postsynaptic potentials (EPSPs) observed in mNTS cells. The membrane models for both the terminal bouton and the mNTS neuron are coupled to separate lumped fluid compartment models describing intracellular Ca2+ ion concentration dynamics. 3. Our modeling strategy is twofold. The first is to validate model performance by reproducing a wide variety of experimental data both from our laboratory and from the literature. The second is to explore the functional aspects of the model in order to gain a greater appreciation for the balance between presynaptic mechanisms (e.g., terminal membrane properties and vesicular dynamics) and postsynaptic mechanisms (e.g., non-NMDA receptor kinetics and neuronal dynamics) that underlie the afferent synaptic drive of mNTS neurons. 4. The model accurately reproduces EPSP dynamics recorded with the use of a wide range of stimulus protocols. The model can also mirror the unique pattern of graded frequency- and use-dependent reduction in peak EPSP magnitude observed experimentally through 60 s of constant, suprathreshold synaptic activation. We demonstrate how vesicular mobilization, recycling, and receptor kinetics can function synergistically in establishing synaptic transfer. Furthermore, we show that by allowing the aggregate rate of vesicle mobilization to respond in a use-dependent manner, it is possible to compensate for the attenuating affects of desensitization at elevated rates of stimulation. 5. Our simulations indicate that the low-frequency characteristics of this synapse are dominated by vesicular dynamics, whereas the high-frequency properties arise from a combination of Ca(2+)-dependent vesicular mobilization and the kinetics of the non-NMDA receptor. Desensitization can influence the peak magnitude and decay time of the EPSP, thereby affecting synaptic throughput. However, we demonstrate that, as the time course of neurotransmitter in the synaptic cleft decreases, the influence of desensitization should be somewhat diminished. As a result, the effective bandwidth of the synapse increases and becomes limited by the gating characteristics of the non-NMDA channel. 6. The model also includes a neuromodulatory aspect in that the frequency response of the synapse can be modulated by an adenylate cyclase-mediated regulatory mechanism. Although our simulations indicate the behavior of a limited number of possible neuromodulatory agents, the results demonstrate the pivotal role such agents could play in modifying synaptic transfer characteristics presynaptically. 7. Both continuous and burst-mode tract stimulation evoke patterns of action potentials in spontaneously active mNTS neurons that are mimicked very well by our model. Our simulations demonstrate that, as the rate of stimulation increases beyond approximately 20-30 Hz, the inherent low-pass frequency-response characteristics of the synapse limit the overall dynamic range of the mNTS neuron, causing the postsynaptic cell to "entrain" at frequencies within its normal operating range.

Afferent Pathways↗

PubMed1994

Identification of dynamic myoelectric signal-to-force models during isometric lumbar muscle contractions.

A 14-muscle myoelectric signal (MES)-driven muscle force prediction model of the L3-L4 cross section is developed which includes a dynamic MES-force relationship and allows for cocontraction. Model parameters are estimated from MES and moments data recorded during rapid exertions in trunk flexion, extension, lateral bending and axial twist. Nine young healthy males participated in the experimental testing. The model used in the parameter estimation is of the output error type. Consistent and physically feasible parameter estimates were obtained by normalizing the RMS MES to maximum exertion levels and using nonlinear constrained optimization to minimize a cost function consisting of the trace of the output error covariance matrix. Model performance was evaluated by comparing measured and MES-predicted moments over a series of slow and rapid exertions. Moment prediction errors were on the order of 25, 30 and 40% during attempted trunk flexion-extensions, lateral bends and axial twists, respectively. The model and parameter estimation methods developed provide a means to estimate lumbar muscle and spine loads, as well as to empirically investigate the use and effects of cocontraction during physical task performances.

Adult↗

PubMed1998

Blood lead slope factor models for adults: comparisons of observations and predictions.

Here we explore the appropriateness of various parameter values for the Bowers et al. model [Risk Anal 14:183-189, 1994] in the context of predicting the influence of site-related exposure to lead in soil on the blood lead (PbB) levels of women of childbearing age. We outline the parameters prescribed by Bowers et al. as well as those prescribed by the U.S. Environmental Protection Agency (U.S. EPA). Comparison of the PbB levels predicted by the Bowers et al. model to those predicted by the validated O'Flaherty pharmacokinetic model indicates that the Bowers et al. model performs favorably when parameter values prescribed here are used. Use of the U.S. EPA-prescribed parameters yields predicted PbB levels that substantially exceed the validated O'Flaherty model predictions. Finally, both the U.S. EPA-prescribed parameter values and the parameter values recommended herein are used to predict PbB levels among adults living in four Superfund communities. Comparison of predicted PbB levels for these communities indicates that the U.S. EPA parameters overstate the incremental influence of lead in soil on PbB levels. Differences between the parameter values prescribed here and the U.S. EPA-prescribed parameters yield substantially different cleanup criteria for lead in soil, although conservative parameter values may still be appropriate for screening purposes.

Adult↗

PubMed1995

A model for global symmetry detection in dense images.

In this paper, a model is proposed for bilateral symmetry detection in images consisting of dense arrangements of local features. The model is elaborated on the basis of a psychophysical experiment showing that grouping precedes and facilitates symmetry detection. The proposed computational model consists of three stages: a grouping stage, a symmetry-detection stage, and a symmetry-subsumption stage. Reliance upon a preliminary grouping stage enables a significant reduction of the computational load for detecting symmetry. An implementation of the model is described, and results are presented, showing a good agreement of the model performance with human symmetry perception.

Computer Simulation↗

PubMed1996

Validation of a methodology for modelling PCDD and PCDF intake via the foodchain.

The US EPA foodchain exposure methodology has been assessed, linking background concentrations of polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) in the atmosphere with eventual intake of these chemicals by humans via the foodchain. The methodology is assessed against background environmental concentrations of PCDD/Fs and the background daily adult intake of PCDD/Fs in the UK diet comprising meats, vegetables, dairy produce, fish and miscellaneous food groups. The environmental fate and transport of PCDD/Fs is estimated for each of the seventeen, 2,3,7,8-positional PCDD/F isomers individually from the atmospheric burden in the atmosphere and into the environmental medium or food product of interest. The model predicts PCDD/F deposition rate and soil concentration to within +/-50% of the measured I-TEQ values. For grass, the model underpredicts PCDD/F concentrations but the lack of a reliable and coherent data set precludes further investigation as to the likely causes. The model performs well on food products, with the exception of potatoes (under-predicts by a factor of 20) and 'other' vegetables (underpredicts by a factor of 3). The total modelled PCDD/F intake via the diet is 61 pg I-TEQ day-1 as opposed to the measured intakes of 69 pg I-TEQ day-1 (excluding plant foods) and 81 pg day-1 (including data for plant foods measured in a previous survey, and not representative of contemporary exposures). The major contributor to total PCDD/F intake is via milk and milk products, accounting for about 40% of the daily dietary intake. 'Hidden' fats and oils account for 35% of the PCDD/F intake, followed by the ingestion meat and meat products (12%). As a group, vegetables do not appear to contribute significantly to the total adult background intake of PCDD/Fs via the diet.

Adult↗

PubMed1997

Testing the self-efficacy-performance linkage of social-cognitive theory.

Past empirical research examining the relationship of self-efficacy perceptions and performance has had several limitations. Most studies were performed in the laboratory with tasks not directly related to individual work performance. As a consequence, many findings are not generalizable to individual work performance. This study tested the self-efficacy-performance model found in Bandura's social-cognitive theory in a work setting, with a sample of 776 American university employees, and with discriminant function analyses. Respondents indicated that performance with computers significantly predicted perceptions of high and low self-efficacy. Results provide additional support for social-cognitive theory as outlined by Bandura.

Adult↗

PubMed1996

Predicting delirium in elderly patients: development and validation of a risk-stratification model.

Delirium is a common and serious complication of acute illness in elderly patients. The aim of this study was to develop and validate a model for predicting development of delirium in elderly medical inpatients who did not have delirium on admission. Consecutive admissions to an acute geriatric unit underwent standardized cognitive assessment every 48 hours. Delirium was diagnosed according to DSM-3 criteria. Independent predictors of delirium in a derivation group of 100 patients were determined using stepwise logistic regression analysis; the predictive model comprised dementia, severe illness and elevated serum urea. This model performed well in a validation group of 84 patients. We conclude that elderly medical patients can be stratified according to their risk for developing delirium using a simple clinical model.

Aged↗

PubMed1992

Analytic sensitivity and specificity of enzyme immunoassay results in testing for human immunodeficiency virus type 1 antibody.

In 1986, a performance evaluation program at the Centers for Disease Control was implemented to assess the quality of performance of laboratories testing for human immunodeficiency virus type 1 antibody and to identify problems that occur during the testing process. Laboratories participating in the Centers for Disease Control Model Performance Evaluation Program for human immunodeficiency virus type 1 antibody testing furnished enzyme immunoassay results after they tested performance evaluation panels that were sent to them in August and November 1989. The panels consisted of 10 individual samples containing antibody-negative and antibody-positive samples, some of which were duplicates. Not all laboratories received the same panel of samples. Low false-negative and false-positive rates, as well as high intrashipment and intershipment reproducibility, indicate that most laboratories did not experience difficulty in testing performance evaluation samples sent to them in August and November 1989.

Evaluation Studies as Topic↗