Network Growth Modeling to Capture Individual Lexical Learning

Table 3

Evaluation of model performance on test data.

Model		Snap. llk	value		AUC	Accuracy	Precision.	Recall
McRae	133	0.478	0.018	34.1	0.689	0.609	0.225	0.702
M. CDI	133	0.516		32.5	0.738	0.679	0.220	0.619
Nelson	534	0.531	0.030	37.9	0.761	0.677	0.256	0.726
N. CDI	534	0.561		38.5	0.764	0.691	0.270	0.668
Phono	677	0.460	0.008	34.6	0.734	0.690	0.238	0.627
P. CDI	677	0.502		31.6	0.767	0.697	0.207	0.664

Note. We include the size of the network representation (or number of words the model makes predictions for), the average negative log-likelihood for each unseen snapshot. We also report the value of an F-test for nested models comparing the CDI-only model to the model containing word importance based on the network growth process. We include the overlap between the words learned by the child and the top words as predicted by the model or percent overlap (). Considering all predictions, we report the area under the curve (AUC) using ROC, accuracy, precision, and recall.