Table 7.

Performance evaluation of the proposed model while considering the user-distribution on Waseem and Hovy dataset. We have restricted the number of tweets per user to 250. The first row for each method represents the Micro average, and the second represents the Macro average.

Method	F1	Prec.	Rec.
Proposed model	81.54	82.21	81.38
	74.09	72.69	76.32
Badjatiya et al. (2017)	82.13	82.16	82.96
	74.76	77.89	73.01
Agrawal and Awekar (2018)	73.89	77.71	72.39
	67.11	66.63	70.76