Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Machine learning model”

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 829 records · Page 46Linked to original sources
PubMed2006

A probabilistic Classifier System and its application in data mining.

The article is about a new Classifier System framework for classification tasks called BYP-CS (for BaYesian Predictive Classifier System). The proposed CS approach abandons the focus on high accuracy and addresses a well-posed Data Mining goal, namely, that of uncovering the low-uncertainty patterns of dependence that manifest often in the data. To attain this goal, BYP-CS uses a fair amount of probabilistic machinery, which brings its representation language closer to other related methods of interest in statistics and machine learning. On the practical side, the new algorithm is seen to yield stable learning of compact populations, and these still maintain a respectable amount of predictive power. Furthermore, the emerging rules self-organize in interesting ways, sometimes providing unexpected solutions to certain benchmark problems.

Algorithms↗
PubMed2003

Learning optimized features for hierarchical models of invariant object recognition.

There is an ongoing debate over the capabilities of hierarchical neural feedforward architectures for performing real-world invariant object recognition. Although a variety of hierarchical models exists, appropriate supervised and unsupervised learning methods are still an issue of intense research. We propose a feedforward model for recognition that shares components like weight sharing, pooling stages, and competitive nonlinearities with earlier approaches but focuses on new methods for learning optimal feature-detecting cells in intermediate stages of the hierarchical network. We show that principles of sparse coding, which were previously mostly applied to the initial feature detection stages, can also be employed to obtain optimized intermediate complex features. We suggest a new approach to optimize the learning of sparse features under the constraints of a weight-sharing or convolutional architecture that uses pooling operations to achieve gradual invariance in the feature hierarchy. The approach explicitly enforces symmetry constraints like translation invariance on the feature set. This leads to a dimension reduction in the search space of optimal features and allows determining more efficiently the basis representatives, which achieve a sparse decomposition of the input. We analyze the quality of the learned feature representation by investigating the recognition performance of the resulting hierarchical network on object and face databases. We show that a hierarchy with features learned on a single object data set can also be applied to face recognition without parameter changes and is competitive with other recent machine learning recognition approaches. To investigate the effect of the interplay between sparse coding and processing nonlinearities, we also consider alternative feedforward pooling nonlinearities such as presynaptic maximum selection and sum-of-squares integration. The comparison shows that a combination of strong competitive nonlinearities with sparse coding offers the best recognition performance in the difficult scenario of segmentation-free recognition in cluttered surround. We demonstrate that for both learning and recognition, a precise segmentation of the objects is not necessary.

Learning↗
PubMed2005

Improving the reliability of medical software by predicting the dangerous software modules.

Software reliability analysis is inevitable for modern medical systems, since a large amount of medical system functionality is now dependent on software, and software does contribute to system failures. Most software reliability models are based on software failure data collected from the project. This creates a problem for the designers since, during the early stage, software failure data are not available. However, a valuable knowledge can be learned from the analysis of previous projects and applied to the new ones. This paper presents the approach that predicts the potentially dangerous software modules under development based on the analysis of the already finished modules using the machine-learning techniques. On the basis of the prediction given by our method software designers are able to devote more testing effort to the dangerous parts of the system, which results in a more reliable medical software system.

Algorithms↗
PubMed2005

Bayesian neural network approaches to ovarian cancer identification from high-resolution mass spectrometry data.

MOTIVATION: The classification of high-dimensional data is always a challenge to statistical machine learning. We propose a novel method named shallow feature selection that assigns each feature a probability of being selected based on the structure of training data itself. Independent of particular classifiers, the high dimension of biodata can be fleetly reduced to an applicable case for consequential processing. Moreover, to improve both efficiency and performance of classification, these prior probabilities are further used to specify the distributions of top-level hyperparameters in hierarchical models of Bayesian neural network (BNN), as well as the parameters in Gaussian process models. RESULTS: Three BNN approaches were derived and then applied to identify ovarian cancer from NCI's high-resolution mass spectrometry data, which yielded an excellent performance in 1000 independent k-fold cross validations (k = 2,...,10). For instance, indices of average sensitivity and specificity of 98.56 and 98.42%, respectively, were achieved in the 2-fold cross validations. Furthermore, only one control and one cancer were misclassified in the leave-one-out cross validation. Some other popular classifiers were also tested for comparison. AVAILABILITY: The programs implemented in MatLab, R and Neal's fbm.2004-11-10.

Bayes Theorem↗
PubMed2025

Artificial intelligence in healthcare and medicine: clinical applications, therapeutic advances, and future perspectives.

Healthcare systems worldwide face growing challenges, including rising costs, workforce shortages, and disparities in access and quality, particularly in low- and middle-income countries. Artificial intelligence (AI) has emerged as a transformative tool capable of addressing these issues by enhancing diagnostics, treatment planning, patient monitoring, and healthcare efficiency. AI's role in modern medicine spans disease detection, personalized care, drug discovery, predictive analytics, telemedicine, and wearable health technologies. Leveraging machine learning and deep learning, AI can analyze complex data sets, including electronic health records, medical imaging, and genomic profiles, to identify patterns, predict disease progression, and recommend optimized treatment strategies. AI also has the potential to promote equity by enabling cost-effective, resource-efficient solutions in low-resource and remote settings, such as mobile diagnostics, wearable biosensors, and lightweight algorithms. Successful deployment requires addressing critical challenges, including data privacy, algorithmic bias, model interpretability, regulatory oversight, and maintaining human clinical oversight. Emphasizing scalable, ethical, and evidence-driven implementation, key strategies include clinician training in AI literacy, adoption of resource efficient tools, global collaboration, and robust regulatory frameworks to ensure transparency, safety, and accountability. By complementing rather than replacing healthcare professionals, AI can reduce errors, optimize resources, improve patient outcomes, and expand access to quality care. This review emphasizes the responsible integration of AI as a powerful catalyst for innovation, sustainability, and equity in healthcare delivery worldwide.

Humans↗
PubMed2025

Foundations of Artificial Intelligence in Hepatology: What a Clinician Needs to Know.

This review focuses on foundational knowledge about artificial intelligence (AI) in hepatology, exploring how AI, including machine learning and deep learning, leverages large-scale clinical data to transform the diagnosis, risk assessment, prognostication, and management of liver diseases. Online resources are described to offer fundamental AI knowledge and essential technical skills and to facilitate clinician participation across the entire AI lifecycle, ensuring they contribute not only as end users but also in development and deployment. Unlike traditional statistical approaches that prioritize interpretable parameters and clinical insight, AI focuses on maximizing predictive accuracy by identifying complex, often non-linear patterns using high-dimensional data, albeit often at the cost of model interpretability. AI is demonstrating clinical utility in liver histopathology and radiological imaging, significantly improving detection accuracy for cirrhosis, clinically significant portal hypertension, and hepatocellular carcinoma. Beyond diagnostics, AI-driven prediction models are emerging to provide personalized risk stratification for the development of liver-related complications and treatment guidance, based on complex data including longitudinal laboratory results, comorbidities, and co-medication use to monitor disease progression and therapy response. The field is rapidly expanding into novel areas such as analyzing patient-reported outcomes, genomic data, and real-time liver function monitoring, offering deeper mechanistic insights alongside clinical tools. Despite the potential to revolutionize hepatology practice and research, successful integration into routine care faces challenges. These include seamless workflow integration with existing electronic health records, establishing clear liability frameworks, and guaranteeing protection of patient privacy. Addressing these hurdles requires collaborative efforts from clinicians, researchers, and regulators to develop best practices and governance. Understanding the transformative capabilities, current applications, emerging frontiers, and essential implementation considerations is crucial for clinicians navigating the evolving AI landscape and responsibly utilizing its power for improved patient outcomes.

PROBAST+AI↗
PubMedDate not supplied

An efficient in silico screening method based on the protein-compound affinity matrix and its application to the design of a focused library for cytochrome P450 (CYP) ligands.

A new method has been developed to design a focused library based on available active compounds using protein-compound docking simulations. This method was applied to the design of a focused library for cytochrome P450 (CYP) ligands, not only to distinguish CYP ligands from other compounds but also to identify the putative ligands for a particular CYP. Principal component analysis (PCA) was applied to the protein-compound affinity matrix, which was obtained by thorough docking calculations between a large set of protein pockets and chemical compounds. Each compound was depicted as a point in the PCA space. Compounds that were close to the known active compounds were selected as candidate hit compounds. A machine-learning technique optimized the docking scores of the protein-compound affinity matrix to maximize the database enrichment of the known active compounds, providing an optimized focused library.

Artificial Intelligence↗
PubMed2006

Exploring predictive and reproducible modeling with the single-subject FIAC dataset.

Predictive modeling of functional magnetic resonance imaging (fMRI) has the potential to expand the amount of information extracted and to enhance our understanding of brain systems by predicting brain states, rather than emphasizing the standard spatial mapping. Based on the block datasets of Functional Imaging Analysis Contest (FIAC) Subject 3, we demonstrate the potential and pitfalls of predictive modeling in fMRI analysis by investigating the performance of five models (linear discriminant analysis, logistic regression, linear support vector machine, Gaussian naive Bayes, and a variant) as a function of preprocessing steps and feature selection methods. We found that: (1) independent of the model, temporal detrending and feature selection assisted in building a more accurate predictive model; (2) the linear support vector machine and logistic regression often performed better than either of the Gaussian naive Bayes models in terms of the optimal prediction accuracy; and (3) the optimal prediction accuracy obtained in a feature space using principal components was typically lower than that obtained in a voxel space, given the same model and same preprocessing. We show that due to the existence of artifacts from different sources, high prediction accuracy alone does not guarantee that a classifier is learning a pattern of brain activity that might be usefully visualized, although cross-validation methods do provide fairly unbiased estimates of true prediction accuracy. The trade-off between the prediction accuracy and the reproducibility of the spatial pattern should be carefully considered in predictive modeling of fMRI. We suggest that unless the experimental goal is brain-state classification of new scans on well-defined spatial features, prediction alone should not be used as an optimization procedure in fMRI data analysis.

Artifacts↗
PubMed2006

Enhancing text categorization with semantic-enriched representation and training data augmentation.

OBJECTIVE: Acquiring and representing biomedical knowledge is an increasingly important component of contemporary bioinformatics. A critical step of the process is to identify and retrieve relevant documents among the vast volume of modern biomedical literature efficiently. In the real world, many information retrieval tasks are difficult because of high data dimensionality and the lack of annotated examples to train a retrieval algorithm. Under such a scenario, the performance of information retrieval algorithms is often unsatisfactory, therefore improvements are needed. DESIGN: We studied two approaches that enhance the text categorization performance on sparse and high data dimensionality: (1) semantic-preserving dimension reduction by representing text with semantic-enriched features; and (2) augmenting training data with semi-supervised learning. A probabilistic topic model was applied to extract major semantic topics from a corpus of text of interest. The representation of documents was projected from the high-dimensional vocabulary space onto a semantic topic space with reduced dimensionality. A semi-supervised learning algorithm based on graph theory was applied to identify potential positive training cases, which were further used to augment training data. The effects of data transformation and augmentation on text categorization by support vector machine (SVM) were evaluated. RESULTS AND CONCLUSION: Semantic-enriched data transformation and the pseudo-positive-cases augmented training data enhance the efficiency and performance of text categorization by SVM.

Algorithms↗
PubMed2005

Extraction and visualization of potential pharmacophore points using support vector machines: application to ligand-based virtual screening for COX-2 inhibitors.

Support vector machines (SVM) were trained to predict cyclooxygenase 2 (COX-2) and thrombin inhibitors. The classifiers were obtained using sets of known COX-2 and thrombin inhibitors as "positive examples" and a large collection of screening compounds as "negative examples". Molecules were encoded by topological pharmacophore-point triangles. In retrospective virtual screening, 50-90% of the known active compounds were listed within the first 0.1% of the ranked database. To check the validity of the constructed classifiers, we developed a method for feature extraction and visualization using SVM. As a result, potential pharmacophore points were weighted according to their importance for COX-2 and thrombin inhibition. Known thrombin and COX-2 pharmacophore points were correctly recognized by the machine learning system. In a prospective virtual screening study, several potential COX-2 inhibitors were predicted and tested in a cellular activity assay. A benzimidazole derivative exhibited significant inhibitory activity with an IC(50) of 0.2 microM, which is better than Celecoxib in our assay. It was demonstrated that the SVM machine-learning method can be used in virtual screening and be analyzed in a human-interpretable way that results in a set of rules for designing novel molecules.

Binding Sites↗
PubMed2025

Transcriptome-wide analysis reveals potential roles of CFD and ANGPTL4 in fibroblasts regulating B cell lineage for extracellular matrix-driven clustering and novel avenues for immunotherapy in breast cancer.

BACKGROUND: The remodeling of the extracellular matrix (ECM) plays a pivotal role in tumor progression and drug resistance. However, the compositional patterns of ECM in breast cancer and their underlying biological functions remain elusive. METHODS: Transcriptome and genome data of breast cancer patients from TCGA database was downloaded. Patients were classified into different clusters by using non-negative matrix factorization (NMF) based on signatures of ECM components and regulators. Weighted Gene Co-expression Network Analysis (WGCNA) was used to identify core genes related to ECM clusters. Additional 10 independent public cohorts including Metabric, SCAN_B, GSE12276, GSE16446, GSE19615, GSE20685, GSE21653, GSE58644, GSE58812, and GSE88770 were collected to construct Training or Testing cohort, following machine learning calculating ECM correlated index (ECI) for survival analysis. Pathway enrichment and correlation analysis were used to explore the relationship among ECM clusters, ECI and TME. Single-cell transcriptome data from GSE161529 was processed for uncovering the differences among ECM clusters. RESULTS: Using NMF, we identified three ECM clusters in the TCGA database: C1 (Neuron), C2 (ECM), and C3 (Immune). Subsequently, WGCNA was employed to pinpoint cluster-specific genes and develop a prognostic model. This model demonstrated robust predictive power for breast cancer patient survival in both the Training cohort (n = 5,392, AUC = 0.861) and the Testing cohort (n = 1,344, AUC = 0.711). Upon analyzing the tumor microenvironment (TME), we discovered that fibroblasts and B cell lineage were the core cell types associated with the ECM cluster phenotypes. Single-cell RNA sequencing data further revealed that angiopoietin like 4 (ANGPTL4)+ fibroblasts were specifically linked to the C2 phenotype, while complement factor D (CFD)+ fibroblasts characterized the other ECM clusters. CellChat analysis indicated that ANGPTL4+ and CFD+ fibroblasts regulate B cell lineage via distinct signaling pathways. Additionally, analysis using the Kaplan-Meier Plotter website showed that CFD was favorable for immunotherapy response, whereas ANGPTL4 negatively impacted the outcomes of cancer patients receiving immunotherapy. CONCLUSION: We identified distinct ECM clusters in breast cancer patients, irrespective of molecular subtypes. Additionally, we constructed an effective prognostic model based on these ECM clusters and recognized ANGPTL4+ and CFD+ fibroblasts as potential biomarkers for immunotherapy in breast cancer.

Humans↗
PubMed2006

Combination of multimodal imaging and molecular genetic information to investigate complex psychiatric disorders.

Multimodal imaging, the combination of several brain imaging techniques in one subject, provides a wealth of parameters and favours the interpretation of complex models in schizophrenia research. Moreover, new imaging tools allow the investigation of distinct neurotransmitter systems and their modulation by pharmacological intervention. An important feature of multimodal imaging is the possibility to characterize the activation dependencies of different neurotransmitters and provide the experimental tool to test system models of brain function and dysfunction. The combination of measurement techniques with high temporal resolution (e. g. MEG, EEG) and high spatial resolution (e. g. fMRI) facilitate the understanding of local and global systems as well as time characteristics. Moreover, the association of imaging parameters with genetic variations of neurotransmitter systems allows the investigation of neurotransmitter activity and its role in the pathophysiology of schizophrenia. To overcome the limitations of standard statistical methods, new approaches in machine learning have to be adapted to handle multiple parameters obtained from brain imaging and genetic measurements.

Brain↗
PubMed2025

Common to rare transfer learning (CORAL) enables inference and prediction for a quarter million rare Malagasy arthropods.

DNA-based biodiversity surveys result in massive-scale data, including up to millions of species-of which, most are rare. Making the most of such data for inference and prediction requires modeling approaches that can relate species occurrences to environmental and spatial predictors, while incorporating information about their taxonomic or phylogenetic placement. Even if the scalability of joint species distribution models to large communities has greatly advanced, incorporating hundreds of thousands of species has not been feasible to date, leading to compromised analyses. Here we present a 'common to rare transfer learning' (CORAL) approach, based on borrowing information from the common species to enable statistically and computationally efficient modeling of both common and rare species. We illustrate that CORAL leads to much improved prediction and inference in the context of DNA metabarcoding data from Madagascar, comprising 255,188 arthropod species detected in 2,874 samples.

Animals↗
PubMed2003

On selecting features from splice junctions: an analysis using information theoretic and machine learning approaches.

The computational recognition of precise splice junctions is a challenge faced in the analysis of newly sequenced genomes. This is challenging due to the fact that the distribution of sequence patterns in these regions is not always distinct. Our objective is to understand the sequence signatures at the splice junctions, not simply to create an artificial recognition system. We use a combination of a neural network based calliper randomization approach and an information theoretic based feature selection approach for this purpose. This has been done in an effort to understand regions that harbor information content and to extract features relevant for the prediction of splice junctions. The analysis using the neural network based calliper randomization approach revealed regions important in the internal representation of the network model. The calliper approach captured both correlated as well as independently important features. The feature selection approach captures features that are independently informative. The two different methods can capture features with different properties. Comparative analysis of the results using both the methods help to infer about the kind of information present in the region.

Alternative Splicing↗
PubMed2025

Radiogenomic MRI biomarkers for noninvasive prediction of GPC3 expression and tumor microenvironment in hepatocellular carcinoma.

BACKGROUND: Glypican-3 (GPC3) is frequently overexpressed in hepatocellular carcinoma (HCC) and plays a key role in immune and metabolic remodeling of the tumor microenvironment. Reliable noninvasive biomarkers for predicting GPC3 status could improve patient stratification and support precision immunotherapy. METHODS: This multicenter retrospective study included 274 patients with pathologically confirmed hepatocellular carcinoma from three institutions, 34 external cases with MRI from The Cancer Imaging Archive, and 363 transcriptomic profiles from The Cancer Genome Atlas. Contrast-enhanced T1-weighted imaging and diffusion-weighted imaging were analyzed. Tumor and peritumoral regions were segmented manually and radiomic features extracted using PyRadiomics. Feature selection was performed with correlation filtering and least absolute shrinkage and selection operator regression. Machine learning classifiers including logistic regression, random forest, support vector machine, k-nearest neighbor, and decision tree were trained with 10-fold cross-validation and tested on independent external cohorts. A radiomics score was calculated for each patient. Radiogenomic analysis correlated radiomics scores with transcriptomic data using weighted gene co-expression network analysis. Hub genes and enriched pathways were identified, and immune infiltration and predicted immunotherapy response were assessed using computational methods. RESULTS: The random forest model using contrast-enhanced T1-weighted imaging achieved an area under the curve of 0.966 in training and 0.935 in internal validation. The integrated contrast-enhanced T1-weighted imaging plus diffusion-weighted imaging model reached an internal validation area under the curve of 0.979. In external testing, the best performance was obtained with a support vector machine model (area under the curve 0.756). Radiomics scores were significantly correlated with GPC3 expression (R&#x2009;=&#x2009;0.78, p&#x2009;<&#x2009;0.05). Transcriptomic analysis identified a 10-gene signature enriched in hypoxia and lipid metabolism pathways that stratified patients into prognostic subgroups (concordance index 0.720, hazard ratio 4.07, p&#x2009;<&#x2009;0.0001). High-risk patients had greater immune infiltration and a lower predicted immune evasion score, suggesting a potential benefit from immunotherapy. CONCLUSIONS: MRI-based radiomics models can noninvasively predict GPC3 expression in hepatocellular carcinoma. Radiomics scores reflect underlying hypoxia and lipid metabolism pathways and stratify patients by prognosis and predicted immunotherapy response. These findings support radiogenomics as a translational approach to imaging-guided precision treatment in hepatocellular carcinoma.

Humans↗
PubMed2025

Liquid Biopsy-Multiomics Link Adhesion Pathway Dysregulation to Kidney Injury Severity.

INTRODUCTION: Severe acute kidney injury (AKI) is strongly associated with the risk of developing chronic kidney disease; however, little is known about the cell type-specific mechanisms driving kidney injury severity. METHODS: In this multicenter observational study, we used clinically obtained liquid biopsy proteomics and machine learning (ML) to predict severe outcomes in patients with COVID-associated and non-COVID AKI. Further, we orthogonally combined 169 urine proteomics with 437 plasma proteomics samples and 40 urine sediment single-cell transcriptomics samples to identify complementary dysregulated mechanisms. RESULTS: Using a 10-fold cross-validated random forest algorithm, we identified a set of urinary proteins that demonstrate predictive power for both discovery and validation set with AUC of 87% and 76%, respectively. These predictive proteomics features obtained demonstrate that cell adhesion and autophagy-associated pathways are uniquely impacted in severe AKI. Differentially abundant proteins (DAPSs) associated with these pathways are highly expressed in cells of the juxtamedullary nephron, endothelial cells (ECs), and podocytes, indicating that these kidney cell types could be potential targets. Single-cell transcriptomic analysis in the in vitro model of kidney organoids infected with SARS-CoV-2 reveal dysregulation of extracellular matrix (ECM) organization in multiple nephron segments, recapitulating the clinically observed fibrotic response across multiomics datasets. Ligand-receptor interaction analysis of the podocyte and tubule organoid clusters shows significant reduction and loss of interaction between integrins and basement membrane receptors in the infected kidney organoids. CONCLUSION: Collectively, these data suggest that ECM degradation and adhesion-associated mechanisms could be the main driver of severe kidney injury.

AKI↗
PubMed2005

Classifying "kinase inhibitor-likeness" by using machine-learning methods.

By using an in-house data set of small-molecule structures, encoded by Ghose-Crippen parameters, several machine learning techniques were applied to distinguish between kinase inhibitors and other molecules with no reported activity on any protein kinase. All four approaches pursued--support-vector machines (SVM), artificial neural networks (ANN), k nearest neighbor classification with GA-optimized feature selection (GA/kNN), and recursive partitioning (RP)--proved capable of providing a reasonable discrimination. Nevertheless, substantial differences in performance among the methods were observed. For all techniques tested, the use of a consensus vote of the 13 different models derived improved the quality of the predictions in terms of accuracy, precision, recall, and F1 value. Support-vector machines, followed by the GA/kNN combination, outperformed the other techniques when comparing the average of individual models. By using the respective majority votes, the prediction of neural networks yielded the highest F1 value, followed by SVMs.

Algorithms↗
PubMed2005

Causal protein-signaling networks derived from multiparameter single-cell data.

Machine learning was applied for the automated derivation of causal influences in cellular signaling networks. This derivation relied on the simultaneous measurement of multiple phosphorylated protein and phospholipid components in thousands of individual primary human immune system cells. Perturbing these cells with molecular interventions drove the ordering of connections between pathway components, wherein Bayesian network computational methods automatically elucidated most of the traditionally reported signaling relationships and predicted novel interpathway network causalities, which we verified experimentally. Reconstruction of network models from physiologically relevant primary single cells might be applied to understanding native-state tissue signaling biology, complex drug actions, and dysfunctional signaling in diseased cells.

Algorithms↗