Xia et al.[75] present different approaches to using SVM’s for RS in a CF set-
ting. They explore the use of Smoothing Support Vector Machines (SSVM). They
also introduce a SSVM-based heuristic (SSVMBH) to iteratively estimate missing
elements in the user-item matrix. They compute predictions by creating a classifier
for each user. Their experimental results report best results for the SSVMBH as
compared to both SSVM’s and traditional user-based and item-based CF. Finally, Oku et al. [27] propose the use of Context-Aware Vector Machines (C-SVM) for
context-aware RS. They compare the use of standard SVM, CSVM and an exten-
sion that uses CF as well as C-SVM. Their results show the effectiveness of the
context-aware methods for restaurant recommendations.