Table 2 Performance comparison of various unimodal models on MIMIC-IV, MIMIC-ECG, and MIMIC-Note datasets. Best and Mean±SD denote the best and mean±standard deviation across cross-validation runs.

From: Integrative multimodal hybrid data fusion for mortality prediction

Dataset

Model

Accuracy

AUC

Precision

Recall

F1-score

Best

Mean±SD

Best

Mean±SD

Best

Mean±SD

Best

Mean±SD

Best

Mean±SD

MIMIC

AdaBoost

0.9195

0.9086±0.0067

0.8037

0.7721±0.0160

0.8252

0.7797±0.0288

0.6358

0.5742±0.0310

0.7100

0.6609±0.0268

Decision Tree

0.8723

0.8640±0.0074

0.7719

0.7459±0.0143

0.5893

0.5612±0.0238

0.6299

0.5745±0.0270

0.6003

0.5675±0.0224

Gradient Boosting

0.9292

0.9190±0.0060

0.8021

0.7772±0.0156

0.8922

0.8606±0.0238

0.6179

0.5715±0.0303

0.7302

0.6865±0.0267

KNN

0.8483

0.8389±0.0040

0.5803

0.5700±0.0074

0.5333

0.4544±0.0315

0.2060

0.1799±0.0165

0.2834

0.2572±0.0188

Logistic Regression

0.8454

0.8434±0.0009

0.5152

0.5046±0.0042

0.6000

0.3617±0.1285

0.0387

0.0131±0.0102

0.0714

0.0250±0.0187

Random Forest

0.9227

0.9137±0.0052

0.7739

0.7496±0.0137

0.9171

0.8838±0.0233

0.5582

0.5116±0.0266

0.6913

0.6478±0.0251

SVM

0.1660

0.1617±0.0021

0.5027

0.4966±0.0040

0.1561

0.1544±0.0011

0.9911

0.9824±0.0078

0.2697

0.2669±0.0019

MIMIC-ECG

Ours

0.9077

0.7460±0.0927

0.9483

0.7912±0.0955

0.9091

0.8501±0.0569

0.8928

0.4345±0.1950

0.8787

0.5576±0.1513

MIMIC-Note

Bert

0.8240

0.7210±0.0747

0.8691

0.7867±0.0747

0.8750

0.7705±0.0589

0.4194

0.3532±0.0564

0.5479

0.4819±0.0584

Back to article page