TABLE 4.

Algorithm and parameters used to get the highest F2 for each set of data.

	Title	Abstracts	T + a	T + M
Algorithm	LR	LR	LR	LR
df	4	4	5	4
n-gram	[1, 2]	[1, 3]	[1, 2]	2
matrix	token-count	TFIDF	token-count	token-count
Precision	0.388	0.500	0.514	0.392
Recall	0.707	0.770	0.779	0.734
F1-score	0.495	0.587	0.614	0.502
F2-score	0.601	0.687	0.701	0.616
ROC AUC	0.910	0.938	0.937	0.930