Recommender system plays an important role in many practical applications that help users to deal with information overload and provide personalized recommendations to them. The context in which a choice is made has been recognized as an important factor for recommendation systems. Recently, researchers extend the classical matrix factorization and allows for a generic integration of contextual information by modeling the data as a tensor. However, current tensor factorization methods suffer from the limitation that the computing cost can be very high in practice. In this paper, we propose GALS, a GPU based parallel tensor factorization algorithm, to accelerate the tensor factorization on large data sets to support the efficient context-aware recommendation. Experiments show that the proposed method can achieve 10 times faster than the current tensor factorization methods.
CITATION STYLE
Zou, B., Lan, M., Li, C., Tan, L., & Chen, H. (2014). Context-aware recommendation using GPU based parallel tensor decomposition. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8933, 213–226. https://doi.org/10.1007/978-3-319-14717-8_17
Mendeley helps you to discover research relevant for your work.