Skip to main content
. 2021 Nov 10;11:22012. doi: 10.1038/s41598-021-01050-7

Table 2.

False-positive rate of the AdaBoost algorithm was higher than false-negative rate.

Predicted: typical (%) Predicted: ASD (%)
Actual: typical 83.9 16.1
Actual: ASD 4.8 95.2

Confusion matrix of classification results using response-type-gaze and significant spatial features as input of the AdaBoost algorithm.