Online Doctor Recommendation with Convolutional Neural Network and Sparse Inputs

Table 4

Experimental evaluation of MAP and NDCG in different dimensions of u, using two metrics.

NCF	u = 10	u = 20	u = 30	u = 40	u = 50
MAP	0.1069	0.1039	0.0886	0.0875	0.0869
NDCG@3	0.3788	0.3539	0.3393	0.3049	0.2489
NDCG@5	0.4291	0.4151	0.3698	0.3611	0.2979
NDCG@10	0.4630	0.4456	0.4118	0.4020	0.3161
NDCG@20	0.4571	0.4321	0.4159	0.4127	0.3412
PMF-CNN	u = 10	u = 20	u = 30	u = 40	u = 50
MAP	0.1278	0.1234	0.1072	0.1068	0.1021
NDCG@3	0.4633	0.4339	0.4160	0.3788	0.3042
NDCG@5	0.5072	0.4748	0.4535	0.4152	0.3651
NDCG@10	0.5451	0.5098	0.4908	0.4461	0.3864
NDCG@20	0.5285	0.5079	0.4952	0.4324	0.4178