Xian Wu, Wei Fan, et al.
WWW 2012
A fully distributed system was proposed for cost-effective data mining. The approach achieved higher accuracy than the centralized and partially distributed learning methods. It also incured less training time, without any communication or computational overheads. Experimental results showed that the total benefits were larger for the fully distributed system than for a partially distributed system using meta-learning.
Xian Wu, Wei Fan, et al.
WWW 2012
Wei Fan, Erheng Zhong, et al.
SDM 2010
Joel L. Wolf, Mark S. Squillante, et al.
IEEE Transactions on Knowledge and Data Engineering
Junyi Xie, Jun Yang, et al.
ICDE 2008