통계검정 테이블 표기 예시

 

Matrix of P Values for Pairwise Comparison Using the McNemar Test.

Fiszman, Marcelo & Chapman, Wendy & Aronsky, Dominik & Evans, R Scott & Haug, Peter. (2000). Automatic Detection of Acute Bacterial Pneumonia from Chest X-ray Reports. Journal of the American Medical Informatics Association : JAMIA. 7. 593-604. 10.1136/jamia.2000.0070593. 

 

Results of the McNemar Test for Comparison of Different Cross-Validation Methods in Terms of Their Accuracy in Classifying Image Patches as Benign or Cancerous

Nir, Guy & Karimi, Davood & Goldenberg, Larry & Fazli, Ladan & Skinnider, Brian & Tavassoli, Peyman & Turbin, Dmitry & Villamil, Carlos & Wang, Gang & Thompson, Darby & Black, Peter & Salcudean, Septimiu. (2019). Comparison of Artificial Intelligence Techniques to Evaluate Performance of a Classifier for Automatic Grading of Prostate Cancer From Digitized Histopathologic Images. JAMA Network Open. 2. e190442. 10.1001/jamanetworkopen.2019.0442. 

 

p values of the McNemar's test for comparing model classifications. Values under 0.05 indicate the error distribution from the two compared models are significantly different. (bold text indicates models that were statistically similar).

de Bem, Pablo & de Carvalho Júnior, Osmar & de Carvalho, Osmar Luiz & Gomes, Roberto & Guimarães, Renato. (2020). Performance Analysis of Deep Convolutional Autoencoders with Different Patch Sizes for Change Detection from Burnt Areas. Remote Sensing. 12. 2576. 10.3390/rs12162576. 

 

McNemar test p-values for the pairwise comparison of the classification results achieved by the investigated CNNs. A significance level of α = 0.05 with the Bonferroni-Holm correction method for multiple comparisons was used. Boldface denotes that the null hypothesis can be rejected.

Militello, Carmelo & Rundo, Leonardo & Vitabile, Salvatore & Conti, Vincenzo. (2021). Fingerprint Classification Based on Deep Learning Approaches: Experimental Findings and Comparisons. Symmetry. 13. 750. 10.3390/sym13050750. 

 

McNemar statistics and associated adjusted p-values from each post-hoc pairwise McNemar test on ISS ≥16 classification performance. Differences between all pairwise combinations were found to be statistically significant except for the direct NMT model and indirect FFNN model comparison.

Doshi, Ayush & Hartka, Thomas. (2024). Comparison of Deep Learning Approaches for Conversion of International Classification of Diseases Codes to the Abbreviated Injury Scale. medRxiv : the preprint server for health sciences. 10.1101/2024.03.06.24303847.