- No file added yet -

The ROC-AUC curves for each model.

figure

posted on 2024-10-02, 06:20 authored by Marie Y. Detrait, Stéphanie Warnon, Raphaël Lagasse, Laurent Dumont, Stéphanie De Prophétis, Amandine Hansenne, Juliette Raedemaeker, Valérie Robin, Géraldine Verstraete, Aline Gillain, Nicolas Depasse, Pierre Jacmin, Delphine Pranger

The characteristics of the best setting for the algorithms to which a GridSearchCV was applied: RF: max_depth: 5, n_estimators: 50, max_samples: 0.75, min_samples_splits: 10, and min_samples_leaf: 2; XGBoost: coldsample_bytree:0.8, learning_rate:0.2, max_depth:5, n_estimators:50, and subsample:0.8; and NuSVC: gamma:0.01 and kernel: rbf. All details are provided in the repository on https://github.com/MarieDetrait-MD/PrimaryRefractoryDLBCL.

History

Usage metrics

PLOS ONE

Categories

Keywords

false positive rate extreme gradient boost enabling early intervention diffuse large b ct scan realization appear highly suitable 54 &# 8211 52 &# 8211 51 &# 8211 46 &# 8211 29 &# 8211 support vector machine random forest classifier primary refractory disease disease management due 012 ), comorbidity overall survival rate 64 &# 8211 40 &# 8211 machine learning approach nb categorical classifier first line treatment auc ), accuracy categorical classifier overall survival supervised machine disease characteristics line therapy 96 ),xlink ">xgboost ),various options variables used validation set using kaplan univariate analysis thirty patients thereby providing significant challenge predict outcomes poor prognosis patients diagnosed novel decision multicenter studies meier method medical context median follow learning techniques influencing factors hematology unit first diagnosis economics status december 2022 cell lymphoma broader scale available based 83 %,8 %)42 patients 2 patients 2 cycles 124 patients

Licence

Exports

RefWorks

BibTeX

Ref. manager

Endnote

DataCite

NLM

DC