Skip to main content
. 2016 Jan 11;50:523–548. doi: 10.1007/s10579-015-9330-7

Fig. 8.

Fig. 8

A 500-fold bootstrapping learning curves generated using YamCha: a chunking and b named entity recognition. Training samples range from 25 to 675 records with a step of 25; testing samples are always set to 75 records