Skip to main content
Fig. 2 | BMC Cancer

Fig. 2

From: Machine learning-based improvement of MDS-CBC score brings platelets into the limelight to optimize smear review in the hematology laboratory

Fig. 2

Machine learning-based improvement of the MDS-CBC score. A Results of the random forest classification for MDS diagnosis of several CBC parameters, the variable importance (VIMP) parameter is plotted. B Performance of the different scores on the cohort to classify patients into 2 groups. Using the MDS-CBC score, 290 non-MDS (81.8%) and 158 MDS (94%) patients are correctly classified. With the 3-parameter RF model, 287 non-MDS (80.4%) and 159 MDS (94.6%) are correctly classified. With the 4-parameter RF model, 310 non-MDS (86.8%) and 158 MDS (94%) are correctly classified. With the 5-parameter RF model, 312 non-MDS (87.4%) and 161 MDS (95.8%) are correctly classified. With the 6-parameter RF model, 300 non-MDS (84%) and 161 MDS (95.8%) are correctly classified

Back to article page

BMC Cancer

ISSN: 1471-2407

Contact us