1.Dir_structure_noduls_from_list.ipynb 2.radiomics_with_diagnosed_nodules.ipynb

CSV: DF_Radiomics_noduls_with_diagnose.csv

Random seed: 42, 80:20 stratified split auf labels

testsplit_fcnn_felix.ipynb

created stratified 80/20 stratified split

DF_Radiomics_noduls_with_diagnose_test_data.csv
DF_Radiomics_noduls_with_diagnose_train_data.csv

scaled dataset

DF_Radiomics_noduls_with_diagnose_train_data_scaled.csv
DF_Radiomics_noduls_with_diagnose_test_data_scaled.csv

FCNN on validation set

Accuracy: 87.09677419354838%
Class 0: Sensitivity: 90.0%, Specificity: 95.1219512195122%
Class 1: Sensitivity: 81.25%, Specificity: 93.61702127659575%
Class 2: Sensitivity: 88.46153846153845%, Specificity: 91.66666666666666%