Table 4 A comparative analysis of 6 machine learning algorithms, implemented in a fusion model, to predict Ki-67 expression based on the internal test dateset.

From: Multi-center study: ultrasound-based deep learning features for predicting Ki-67 expression in breast cancer

Model

Acc

AUC

95% CI

Sens

Spec

F1

LR

0.745

0.785

0.723–0.847

0.778

0.679

0.801

SVM

0.778

0.811

0.752–0.870

0.823

0.691

0.831

KNN

0.573

0.645

0.570–0.720

0.532

0.654

0.622

RandomForest

0.665

0.677

0.604–0.750

0.759

0.481

0.750

LightGBM

0.749

0.767

0.700–0.834

0.791

0.667

0.806

AdaBoost

0.732

0.750

0.679–0.822

0.797

0.605

0.797

Back to article page