Skip to main content

Table 1 Averaged number of iteration steps during the boosting in the four classification models (full alternative (full-ADT) or classic boosting decision trees (BDT) paired with exponential (Exp) or log (Log) loss function in the four populations.

From: Boosting alternating decision trees modeling of disease trait information

Model

Aipotu

Danacca

Karangar

NYC

full-ADT & exponential

10.9

9.1

11.8

12.3

full-ADT & log

10.4

7.7

12.7

11.6

BDT & exponential

13.2

12.9

10.6

14.8

BDT & log

14.2

13.9

14

13.9