Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “probabilistic modelling”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 955 records · Page 53Linked to original sources
PubMed1998

Computer-based decision support in the management of primary gastric non-Hodgkin lymphoma.

Primary non-Hodgkin lymphoma of the stomach is a rare disorder for which clinical management has not yet been settled completely. Faced with the many uncertainties associated with the selection of a treatment for a patient with this disorder, it is difficult to determine the treatment that is optimal for the patient, as well as the prognosis to be expected. The development of a decision-theoretic model of non-Hodgkin lymphoma of the stomach is described. The model aims to assist the clinician in exploring various clinical questions, among others questions concerning prognosis and optimal treatment. Central to the model is a probabilistic network that offers an explicit representation of the uncertainties underlying the decision-making process. The model has been incorporated in a decision-support system. Preliminary evaluation results indicate that the performance of the model in its present form matches the performance of experienced clinicians.

Combined Modality Therapy↗
PubMed2004

Automatically parcellating the human cerebral cortex.

We present a technique for automatically assigning a neuroanatomical label to each location on a cortical surface model based on probabilistic information estimated from a manually labeled training set. This procedure incorporates both geometric information derived from the cortical model, and neuroanatomical convention, as found in the training set. The result is a complete labeling of cortical sulci and gyri. Examples are given from two different training sets generated using different neuroanatomical conventions, illustrating the flexibility of the algorithm. The technique is shown to be comparable in accuracy to manual labeling.

Algorithms↗
PubMed1989

Stochastic choice models: A comparison between Bush-Mosteller and a source-independent reward-following model.

Horner and Staddon (1987) argued that a class of reward-following processes defined by a property they termed ratio invariance is a better model for the probabilistic choice performance of pigeons than competing molecular accounts such as momentary maximizing, melioration, and the Bush-Mosteller model. The critical data were provided by choice distributions-distributions of a variable S, the proportion of Right choices, defined on a moving window typically 32 choices long-obtained under a frequency-dependent schedule. The schedule prescribed equal payoff probabilities, p(S), for both choices. p(S) was a maximum when S = 0.5 and declined linearly for S values above and below 0.5. Pigeons showed generally bimodal choice distributions with the modes at equal p(S) values. These data do not follow easily from melioration or momentary maximizing and are inconsistent with molar maximizing, but they may be consistent with Bush-Mosteller. We present here the results of computer simulations showing that the ratio-invariance model studied yields, as expected, choice modes at equal p(S) values, but that Bush-Mosteller, although capable of generating bimodal choice distributions, does not have choice modes at equal p(S) values.

Journal Article↗
PubMed2004

Parallel processing in visual search asymmetry.

The difficulty of visual search may depend on assignment of the same visual elements as targets and distractors-search asymmetry. Easy C-in-O searches and difficult O-in-C searches are often associated with parallel and serial search, respectively. Here, the time course of visual search was measured for both tasks with speed-accuracy methods. The time courses of the 2 tasks were similar and independent of display size. New probabilistic parallel and serial search models and sophisticated-guessing variants made predictions about time course and accuracy of visual search. The probabilistic parallel model provided an excellent account of the data, but the serial model did not. Asymptotic search accuracies and display size effects were consistent with a signal-detection analysis, with lower variance encoding of Cs than Os. In the absence of eye movements, asymmetric visual search, long considered an example of serial deployment of covert attention, is qualitatively and quantitatively consistent with parallel search processes.

Cognition↗
PubMed1999

Total system performance assessment for waste disposal using a logic tree approach.

The Electric Power Research Institute (EPRI) has sponsored the development of a model to assess the long-term, overall "performance" of the candidate spent fuel and high-level radioactive waste (HLW) disposal facility at Yucca Mountain, Nevada. The model simulates the processes that lead to HLW container corrosion, HLW mobilization from the spent fuel, and transport by groundwater, and contaminated groundwater usage by future hypothetical individuals leading to radiation doses to those individuals. The model must incorporate a multitude of complex, coupled processes across a variety of technical disciplines. Furthermore, because of the very long time frames involved in the modeling effort (>> 10(4) years), the relative lack of directly applicable data, and many uncertainties and variabilities in those data, a probabilistic approach to model development was necessary. The developers of the model chose a logic tree approach to represent uncertainties in both conceptual models and model parameter values. The developers felt the logic tree approach was the most appropriate. This paper discusses the value and use of logic trees applied to assessing the uncertainties in HLW disposal, the components of the model, and a few of the results of that model. The paper concludes with a comparison of logic trees and Monte Carlo approaches.

Geological Phenomena↗
PubMed1996

On the interpretation of certainty factors in expert systems.

Despite the strong theoretical foundation the Bayesian probabilistic approach to model uncertainty in medicine meets many difficulties at the implementation step. One of these difficulties is related to a large amount of conditional probabilities to be assessed and in many cases this task was recognised to be practically insoluble. The MYCIN certainty factors model is a widely distributed pragmatical approach for modeling reasoning under uncertainty that substantially simplifies the problem, at the sacrifice of theoretical soundness. One can determine certainty factors as a function of prior and posterior probability. However, this approach is only consistent with the modularity axiom for certainty factors for tree-structure inference networks, which is rarely true for practical applications. In this paper we abandon the requirement of a direct probabilistic interpretation of certainty factors and build a model of propagation of uncertainty in terms of absolute belief and belief updates. We describe our model for propagating uncertainty in terms of matrix multiplication with specifically defined addition and multiplication which correspond to parallel and sequential combinations of certainty factors. It is possible to define these operations in such a manner that they form a field, and therefore to obtain some useful properties. Finally we present a method of determining certainty factors from statistical data using nonlinear regression and illustrate it with a leukemia diagnostics problem.

Artificial Intelligence↗
PubMed2005

Dynamic Trees for unsupervised segmentation and matching of image regions.

We present a probabilistic framework--namely, multiscale generative models known as Dynamic Trees (DT)--for unsupervised image segmentation and subsequent matching of segmented regions in a given set of images. Beyond these novel applications of DTs, we propose important additions for this modeling paradigm. First, we introduce a novel DT architecture, where multilayered observable data are incorporated at all scales of the model. Second, we derive a novel probabilistic inference algorithm for DTs--Structured Variational Approximation (SVA)--which explicitly accounts for the statistical dependence of node positions and model structure in the approximate posterior distribution, thereby relaxing poorly justified independence assumptions in previous work. Finally, we propose a similarity measure for matching dynamic-tree models, representing segmented image regions, across images. Our results for several data sets show that DTs are capable of capturing important component-subcomponent relationships among objects and their parts, and that DTs perform well in segmenting images into plausible pixel clusters. We demonstrate the significantly improved properties of the SVA algorithm--both in terms of substantially faster convergence rates and larger approximate posteriors for the inferred models--when compared with competing inference algorithms. Furthermore, results on unsupervised object recognition demonstrate the viability of the proposed similarity measure for matching dynamic-structure statistical models.

Algorithms↗
PubMed2026

BaGGLS: a Bayesian shrinkage framework for interpretable modeling of interactions in high-dimensional biological data.

MOTIVATION: Biological data is often high dimensional, noisy, and governed by complex interactions among sparse signals. This poses major challenges for interpretability and reliable feature selection. Tasks such as identifying motif interactions in genomics exemplify these difficulties, as only a small subset of biologically relevant features (e.g. motifs) are typically active, and their effects are often non-linear and context-dependent. While statistical approaches often result in more interpretable models, deep learning models have proven effective in modeling complex interactions and prediction accuracy, yet their black-box nature limits interpretability. RESULTS: We introduce BaGGLS, a flexible and interpretable probabilistic binary regression model designed for high-dimensional biological inference involving feature interactions. BaGGLS incorporates a Bayesian group global-local shrinkage prior, aligned with the group structure introduced by interaction terms. This prior encourages sparsity while retaining interpretability, helping to isolate meaningful signals and suppress noise. To enable scalable inference, we employ a partially factorized variational approximation that captures posterior skewness and supports efficient learning even in large feature spaces. In extensive simulations, we compare BaGGLS to frequentist probit regressions (unconstrained and with L1-penalty) as well as a probit model with Markov Chain Monte Carlo (MCMC) sampling under a horseshoe prior. We can show that BaGGLS outperforms the other methods with regard to interaction detection and is many times faster than MCMC sampling under the horseshoe prior. We also demonstrate the usefulness of BaGGLS in the context of interaction discovery from motif scanner outputs (e.g. Find Individual Motif Occurrences (FIMO)) and noisy attribution scores from deep learning models. This shows that BaGGLS is a promising approach for uncovering biologically relevant interaction patterns, with potential applicability across a range of high-dimensional tasks in computational biology. AVAILABILITY: Code is available at gitlab.com/dacs-hpi/baggls.

Bayes Theorem↗
PubMed2006

A jumping profile Hidden Markov Model and applications to recombination sites in HIV and HCV genomes.

BACKGROUND: Jumping alignments have recently been proposed as a strategy to search a given multiple sequence alignment A against a database. Instead of comparing a database sequence S to the multiple alignment or profile as a whole, S is compared and aligned to individual sequences from A. Within this alignment, S can jump between different sequences from A, so different parts of S can be aligned to different sequences from the input multiple alignment. This approach is particularly useful for dealing with recombination events. RESULTS: We developed a jumping profile Hidden Markov Model (jpHMM), a probabilistic generalization of the jumping-alignment approach. Given a partition of the aligned input sequence family into known sequence subtypes, our model can jump between states corresponding to these different subtypes, depending on which subtype is locally most similar to a database sequence. Jumps between different subtypes are indicative of intersubtype recombinations. We applied our method to a large set of genome sequences from human immunodeficiency virus (HIV) and hepatitis C virus (HCV) as well as to simulated recombined genome sequences. CONCLUSION: Our results demonstrate that jumps in our jumping profile HMM often correspond to recombination breakpoints; our approach can therefore be used to detect recombinations in genomic sequences. The recombination breakpoints identified by jpHMM were found to be significantly more accurate than breakpoints defined by traditional methods based on comparing single representative sequences.

Algorithms↗
PubMedDate not supplied

Continuous trees and NEVADA simulation: a quadrature approach to modeling continuous random variables in decision analysis.

This paper introduces an improved technique for modeling risk and decision problems that have continuous random variables and probabilistic dependence. Variables are modeled with mixtures of four-parameter random variables, called "continuous trees." Functions of random variables are calculated using gaussian quadrature in a manner called "Nevada simulation" (NumErical Integration of Variance And probabilistic Dependence Analyzer). This technique is compared with traditional decision-tree modeling in terms of analytic technique, solution-time complexity, and accuracy. Nevada simulation takes advantage of the probabilistic independence in a decision problem while allowing for probabilistic dependence to achieve polynomial computational-time complexity for many decision problems. It improves on the accuracy of traditional decision trees by employing larger approximations than traditional decision analysis. It improves on traditional decision analysis by modeling continuous variables with continuous, rather than discrete, distributions. A Bayesian analysis using a mixed discrete-continuous probability distribution for cigarette smoking rate is presented.

Algorithms↗
PubMed2005

W3MCSim: an online and reconfigurable Monte Carlo simulator for interactive probabilistic/statistical modeling.

We have implemented a Monte Carlo simulator (W3MCSim) as an Internet software tool, primarily for interactive use by students, educators, life and other physical scientists, as well as other practitioners of probabilistic and statistical modeling. Interested users download, install and run W3MCSim by visiting the application website. This application incorporates three freely available Microsoft web technologies, namely the Internet Explorer web browser, the Component Object Model software framework and the JScript web page script interpreter. We define the software architecture here, as a web application model, and show how incorporation of these technologies provides an efficient solution to W3MCSim software deployment. We demonstrate the usability and versatility of this simulator with three distinct tutorial examples: simulating the sum of six-sided dice, estimating intersection frequency in "Buffon's needle problem", and testing an animal experiment design model a priori. We also show how the program components can be reconfigured into other programs.

Computer Simulation↗
PubMed1976

Dynamic decision models for clinical diagnosis.

A unified approach to clinical decision-making is presented. This combines partially observable Markovian decision processes (Markov or semi-Markov) with cause-effect models as a probabilistic representation of the diagnostic process. Pattern recognition techniques are used in a first stage of system state identification. This new class of dynamic models has a direct application to medical diagnosis and treatment and specific physiological examples are emphasised. The methodology is given for combining the patient state of health, the clinician's state of knowledge of the cause-effect representation from the observation space (measurements), feature selection using pattern recognition techniques and, finally, the treatment decisions with which to restore the patient to a more desirable state of health. A cost functional for the decision process has then to be optimised according to some pre-assigned objective function (social return from the patient state of health or treatment cost for the patient), when the process has an infinite time horizon.

Computers↗
PubMed2005

Development of an economic model to assess the cost effectiveness of treatment interventions for chronic obstructive pulmonary disease.

OBJECTIVE: To develop a Markov model that allows the cost effectiveness of interventions in patients with chronic obstructive pulmonary disease (COPD) to be estimated, and to apply the model to investigate the cost effectiveness of an inhaled corticosteroid/long-acting beta(2)-adrenoceptor agonist (beta(2)-agonist) combination (salmeterol/fluticasone propionate) versus usual care. METHODS: A Markov model consisting of four mutually exclusive disease states was constructed (mild, moderate and severe disease, and death). The transition probabilities of disease progression (for smokers and ex-smokers) and death were derived from the published medical literature. The model outputs were costs, exacerbations, survival, QALYs and cost effectiveness. The model was made fully probabilistic to reflect the joint uncertainty in the model parameters. Efficacy data for the combination of inhaled salmeterol/fluticasone propionate 50/500microg twice daily in poorly reversible COPD patients with a history of exacerbations were obtained from the 1-year TRISTAN (TRial of Inhaled STeroids ANd long-acting beta-agonists) study and applied to the model, based on patient profiles representative of COPD clinical trials. RESULTS: According to the model, the mean life expectancy with usual care alone (placebo group) was 8.95 years, which decreased to 4.08 QALYs once adjusted for quality and discounted, at a lifetime discounted cost of Can 16,415 dollars per patient (year 2002 values). Assuming that salmeterol/fluticasone propionate reduced exacerbation frequency only (base case analysis), the estimated mean survival time remained unchanged but there was an increase in the number of QALYs (4.21) for an estimated lifetime cost of Can 25,780 dollars, resulting in a cost-effectiveness ratio of Can 74,887 dollars per QALY (95% CI 21,985, 128,671) versus usual care. If a survival benefit was assumed for salmeterol/fluticasone propionate, the incremental cost per QALY was Can11,125 dollars (95% CI 8710, dominated) versus usual care. If the combination achieved around a 10% improvement in forced expiratory volume in 1 second, leading to delayed progression to more severe disease states, the benefits translated into an incremental cost per QALY of Can 49,928 dollars (95% CI 37 269, 66,006) versus usual care. CONCLUSIONS: This Markov model allows, for the first time, a means of estimating the long-term cost effectiveness and cost utility of interventions for COPD. Initial evidence suggests that for patients with poorly reversible COPD and a documented history of frequent COPD exacerbations, the addition of salmeterol (a long-acting beta(2)-agonist) to fluticasone propionate (an inhaled corticosteroid) is potentially cost effective from the Canadian healthcare payer's perspective. However, the precision of this estimate will be improved when additional data are available from clinical trials such as the ongoing TORCH (TOwards a Revolution in COPD Health) study.

Adrenergic beta-Agonists↗
PubMedDate not supplied

Modeling turbulent diffusion and advection of indoor air contaminants by Markov chains.

Turbulent eddy diffusion models are used to describe a continuous concentration gradient with distance from an in-room contaminant emission source. A refined diffusion model termed the Drivas model also accounts for contaminant reflection by wall surfaces and partially accounts for removal by exhaust air. This article develops two models based on Markov chains to describe indoor air contaminant dispersion by turbulent diffusion and advection, and removal by the exhaust airflow. Markov model I is equivalent to the Drivas model and is computationally simple. Markov model II can provide more realism by accounting for the locations of air inlets and outlets, advective flow patterns, in-room reflective surfaces, and contaminant removal mechanisms at specific room positions. The price paid for this greater realism is greater computational complexity. Both Markov models are explicitly probabilistic and estimate the expected concentration values at given room positions.

Air Pollution, Indoor↗
PubMedDate not supplied

Medical informatics and clinical decision making: the science and the pragmatics.

There are important scientific and pragmatic synergies between the medical decision making field and the emerging discipline of medical informatics. In the 1970s, the field of medicine forced clinically oriented artificial intelligence (AI) researchers to develop ways to manage explicit statements of uncertainty in expert systems. Classic probability theory was considered and discussed, but it tended to be abandoned because of complexities that limited its use. In medical AI systems, uncertainty was handled by a variety of ad hoc models that simulated probabilistic considerations. To illustrate the scientific interactions between the fields, the author describes recent work in his laboratory that has attempted to show that formal normative models based on probability and decision theory can be practically melded with AI methods to deliver effective advisory tools. In addition, the practical needs of decision makers and health policy planners are increasingly necessitating collaborative efforts to develop a computing and communications infrastructure for the decision making and informatics communities. This point is illustrated with an example drawn from outcomes management research.

Artificial Intelligence↗
PubMed2006

Pharmacoeconomic analysis of recombinant factor VIIa versus APCC in the treatment of minor-to-moderate bleeds in hemophilia patients with inhibitors.

OBJECTIVE: To compare the cost-effectiveness of three treatment regimens using recombinant activated Factor VII (rFVIIa), NovoSeven, and activated prothrombin-complex concentrate (APCC), FEIBA VH, for home treatment of minor-to-moderate bleeds in hemophilia patients with inhibitors. METHODS: A literature-based, decision-analytic model was developed to compare three treatment regimens. The regimens consisting of first-, second-, and third-line treatments were: rFVIIa-rFVIIa-rFVIIa; APCC-rFVIIa-rFVIIa; and APCC-APCC-rFVIIa. Patients not responding to first-line treatment were administered second-line treatment, and those failing second-line received third-line treatment. Using literature and expert opinion, the model structure and base-case inputs were adapted to the US from a previously published analysis. The percentage of evaluable bleeds controlled with rFVIIa and APCC were obtained from published literature. Drug costs (2005 US$) based on average wholesale price were included in the base-case model. Univariate and probabilistic sensitivity analyses (second-order Monte Carlo simulation) were conducted by varying the efficacy, re-bleeding rates, patient weight, and dosing to ascertain robustness of the model. RESULTS: In the base-case analysis, the average cost per resolved bleed using rFVIIa as first-, second-, and third-line treatment was $28 076. Using APCC as first-line and rFVIIa as second- and third-line treatment resulted in an average cost per resolved bleed of $30 883, whereas the regimen using APCC as first- and second-line, and rFVIIa as third-line treatment was the most expensive, with an average cost per resolved bleed of $32 150. Cost offsets occurred for the rFVIIa-only regimen through avoidance of second and third lines of treatment. In probabilistic sensitivity analyses, the rFVIIa-only strategy was the least expensive strategy more than 68% of the time. CONCLUSIONS: The management of minor-to-moderate bleeds extends beyond the initial line of treatment, and should include the economic impact of re-bleeding and failures over multiple lines of treatment. In the majority of cases, the rFVIIa-only regimen appears to be a less expensive treatment option in inhibitor patients with minor-to-moderate bleeds over three lines of treatment.

Blood Coagulation Factors↗
PubMed1988

Three-dimensional computer graphic modeling of ballistic injuries.

Multiple variables affect the tissue destruction caused by missiles, and the interaction of these variables is incompletely understood. The recently developed technology of computerized solid modeling now makes it possible to analyze these parameters in three dimensions. A technique for creating solid models of organic structures is described. The tissues within the boundaries thus defined are ascribed physical attributes by means of finite element analysis with data derived from empirical studies. An interactive user-friendly program is being developed combining this modeling with a probabilistic scheme (Monte Carlo simulation) to describe a variety of wounding scenarios. The data from these predictions will be compared with information from wound registries and the model refined until it can project consistently accurate patterns of injury.

Computer Graphics↗
PubMed2001

Probabilistic neural networks using Bayesian decision strategies and a modified Gompertz model for growth phase classification in the batch culture of Bacillus subtilis.

Probabilistic neural networks (PNNs) were used in conjunction with the Gompertz model for bacterial growth to classify the lag, logarithmic, and stationary phases in a batch process. Using the fermentation time and the optical density of diluted cell suspensions, sampled from a culture of Bacillus subtilis, PNNs enabled a reliable determination of the growth phases. Based on a Bayesian decision strategy, the Gompertz based PNN used newly proposed definition of the lag and logarithmic phases to estimate the latent, logarithmic and stationary phases. This network topology has the potential for use with on-line turbidimeter for the automation and control of cultivation processes.

Journal Article↗