2024
An Efficient Contrastive Unimodal Pretraining Method for EHR Time Series Data
King R, Kodali S, Krueger C, Yang T, Mortazavi B. An Efficient Contrastive Unimodal Pretraining Method for EHR Time Series Data. 2024, 00: 1-8. DOI: 10.1109/bhi62660.2024.10913624.Peer-Reviewed Original ResearchDeep neural networksElectronic health recordsMachine learningSelf-supervised taskSelf-supervised learningSemi-supervised learningEffective feature extractionMIMIC-III datasetExtract meaningful informationTimeseries dataExtract valuable insightsSOTA methodsContrastive pretrainingLabeled dataFeature extractionNeural networkData batchesEICU datasetTime series dataMeaningful informationMIMIC-IIILinear evaluationComplex mappingComputational demandsHealth recordsStrokeClassifier: ischemic stroke etiology classification by ensemble consensus modeling using electronic health records
Lee H, Schwamm L, Sansing L, Kamel H, de Havenon A, Turner A, Sheth K, Krishnaswamy S, Brandt C, Zhao H, Krumholz H, Sharma R. StrokeClassifier: ischemic stroke etiology classification by ensemble consensus modeling using electronic health records. Npj Digital Medicine 2024, 7: 130. PMID: 38760474, PMCID: PMC11101464, DOI: 10.1038/s41746-024-01120-w.Peer-Reviewed Original ResearchElectronic health recordsWeighted F1MIMIC-IIIClinical decision support systemsMulti-class classificationNatural language processingMIMIC-III datasetHealth recordsMachine Learning ClassifiersDecision support systemArtificial intelligence toolsVascular neurologistsLearning classifiersBinary classificationCross-validation accuracyLanguage processingMeta-modelIntelligence toolsStroke prevention effortsAcute ischemic strokeStroke etiologySupport systemStroke etiology classificationClassification toolClassifier
2018
Identifying and characterizing highly similar notes in big clinical note datasets
Gabriel R, Kuo T, McAuley J, Hsu C. Identifying and characterizing highly similar notes in big clinical note datasets. Journal Of Biomedical Informatics 2018, 82: 63-69. PMID: 29679685, DOI: 10.1016/j.jbi.2018.04.009.Peer-Reviewed Original ResearchConceptsClinical note datasetsDe-duplication algorithmMIMIC-III datasetElectronic health recordsJaccard similarityDe-duplicationLocality Sensitive HashingMIMIC-IIINear-duplicatesScalable algorithmMeasure similarityAccurate statistical modelsSources of duplicationClustering methodDatasetAlgorithmApproximation algorithmHealth recordsDisjoint setsInstitutional datasetComparison of notesPairs of notesHashPairwise comparisonsPairwise
This site is protected by hCaptcha and its Privacy Policy and Terms of Service apply