Figure 3 | Scientific Reports

Figure 3

From: Prediction of amyloid β PET positivity using machine learning in patients with suspected cerebral amyloid angiopathy markers

Figure 3

Cut-off values of important variables in GBM and RF models. GBM = gradient boosting machines, RF = random forest, CMBs = cerebral microbleeds, In PDP curve (y is threshold of metrics, and x is cut-off value), the optimal cut-off value was determined, when the curve passes the threshold which was obtained above. (A) Cut-off values of variables to predict Aβ positivity in GBM were as follows: (1) If the number of lobar CMB is more than 16.4, (2) if there is no deep CMBs, (3) if the number of lacunes is more than 7.4, (4) if age is older than 74.3 (GBM), (5) if there is no CMBs in dentate nucleus. (B) Cut-off values of variables to predict Aβ positivity in RF were as follows: (1) If the number of lobar CMB is more than 14.3 , (2) if there is no deep CMBs, (3) if the number of lacunes is more than 7.4, (4) if age is older than 64, (5) if there is no CMBs in dentate nucleus.

Back to article page