Table 16.

Comparison of ROUGE-1 recall and F-scores (Twitter-specific tags, emoticons, hashtags, mentions, urls, removed and standard rouge stemming(-m) and stopwords(-s) option) for MEDSUM (the proposed methodology) and the baseline method COWTS for treatment class

Event	MEDSUM		COWTS
	Recall	F-score	Recall	F-score
Ebola	0.4803	0.4621	0.3858	0.3525
MERS	0.6517	0.5983	0.4642	0.4244

For each evaluation metric, the result of better performing system has been boldfaced