Research Article

Mobile Service Recommendation via Combining Enhanced Hierarchical Dirichlet Process and Factorization Machines

Table 3

MAE and RMSE comparison of multiple recommendation approaches.

MethodMatrix density = 10%Matrix density = 20%Matrix density = 30%
MAERMSEMAERMSEMAERMSE

Given 10SPCC0.42580.56430.40050.52570.39320.5036
MPCC0.43160.57010.41080.52930.40350.5113
PMF0.24170.38350.22630.37740.20140.3718
LDA-FMs0.20910.32250.19690.31160.18320.3015
HDP-FMs0.15470.28740.13290.26690.12830.2498
EHDP-FMs0.13080.25070.11540.23720.10810.2093

Given 20SPCC0.41350.55410.39180.51580.38900.5003
MPCC0.44130.57120.42210.52020.41510.5109
PMF0.23980.35590.21370.34270.19920.3348
LDA-FMs0.19890.31040.19070.30180.18010.2894
HDP-FMs0.14860.27130.12970.25130.11850.2291
EHDP-FMs0.12270.24190.10550.22160.09520.1904

Given 30SPCC0.40160.54470.39070.51070.37390.5012
MPCC0.45180.57710.43170.51590.42390.5226
PMF0.22140.33190.20910.31170.19860.3052
LDA-FMs0.19700.30960.18650.29930.17940.2758
HDP-FMs0.13770.25560.11090.24610.10470.2057
EHDP-FMs0.11130.22480.09260.20570.08040.1673