Skip to main content

Table 10 Average correct rates and SDs in classifying chest CT images as COVID-19 positive/negative when Inception-v3 and each algorithm hyperparameter combination in Table 6 were used in five independent experimental runs

From: Classifying chest CT images as COVID-19 positive/negative using a convolutional neural network ensemble model and uniform experimental design method

Model# experiment number

Dataset

Experimental runs

 

1

2

3

4

5

Average

SD

Inception-v3#1

Training set

0.7418

0.7533

0.75

0.75

0.7484

0.7487

0.00425

Validation set

0.7172

0.7475

0.7576

0.7374

0.7374

0.7394

0.01498

Inception-v3#2

Training set

0.6814

0.6699

0.683

0.6716

0.6716

0.6755

0.00618

Validation set

0.7172

0.7071

0.7172

0.7071

0.7071

0.7111

0.00553

Inception-v3#3

Training set

0.9869

0.9869

0.9853

0.9869

0.982

0.9856

0.00213

Validation set

0.8283

0.8485

0.8687

0.8485

0.8485

0.8485

0.01428

Inception-v3#4

Training set

0.5163

0.5196

0.5229

0.5147

0.5147

0.5176

0.00356

Validation set

0.5556

0.5657

0.5859

0.5859

0.5859

0.5758

0.01428

Inception-v3#5

Training set

0.9118

0.8938

0.902

0.9003

0.9003

0.9016

0.00649

Validation set

0.8182

0.8182

0.8283

0.7879

0.798

0.8101

0.0166

Inception-v3#6

Training set

0.8448

0.817

0.8301

0.8513

0.8513

0.8389

0.01499

Validation set

0.798

0.7677

0.7677

0.7879

0.7879

0.7818

0.01355

Inception-v3#7

Training set

0.9869

0.9869

0.9918

0.9935

0.9853

0.9889

0.00355

Validation set

0.8788

0.8788

0.8586

0.8687

0.8485

0.8667

0.01317