Fig. 1.

Confusion matrix and feature importance of the extremely randomized trees classifier. (a) The row of confusion matrix denotes the true label in the dataset, and the column denotes the predicted label of each classifier. Each cell of the matrix indicates the average ratio value. The color of the cell indicates the value of the cell. (b) Importance value of every feature class is represented as a pie chart. Importance values for the same feature class are all summed up. TE, Tae-Eum type; SE, So-Eum type; SY, So-Yang type.