Table 2.

Best and worst parameter values for two- and three- layer Tybalt models for real datasets. l: learning rate, b: batch size, e: epoch, c: dimensionality of the first hidden layer.

2-layer model
	Best combination						Worst combination
Datasets	l	b	e	c	NMI	ARI	l	b	e	c	NMI	ARI
Baron et al.	0.0005	100	25	100	0.64	0.38	0.002	50	200	500	0.36	0.17
Wang et al.	0.001	200	200	500	0.46	0.3	0.0005	200	25	100	0.2	0.11
Camp et al.	0.002	50	25	500	0.81	0.71	0.0005	100	25	100	0.64	0.47
3-layer model
	Best combination						Worst combination
Datasets	l	b	e	c	NMI	ARI	l	b	e	c	NMI	ARI
Baron et al.	0.0005	100	25	500	0.63	0.36	0.001	100	200	500	0.33	0.17
Wang et al.	0.0005	50	200	500	0.45	0.3	0.0005	200	25	100	0.24	0.13
Camp et al.	0.0005	200	50	500	0.76	0.62	0.0005	200	25	250	0.61	0.42