[Retracted] The Design of Critical Care Information System Supporting Clinical Decision Based on Deep Learning Recognition Method

Table 2

The best results of each data set of the user-based collaborative filtering model.

Missing	Raw data	10%	20%	30%
Accuracy of skin disease data set	0.96894 (k = 1)	0.96382 (k = 1)	0.92764 (k = 1)	0.82814 (k = 8)
Skin disease data set F 1-S core	0.96787 (k = 1)	0.96313 (k = 1)	0.8937 (k = 1)	0.7251 (k = 8)
Chronic nephritis accuracy rate	0.97183 (k = 4)	0.96894 (k = 6)	0.97041 (k = 7)	0.95275 (k = 6)
Chronic nephritis F 1-S core	0.97019 (k = 4)	0.96894 (k = 6)	0.96894 (k = 7)	0.95077 (k = 6)