A fast and accurate collaborative filter
Wanyu Deng, Qinghua Zheng, Lin Chen · 2009
There are two key issues for collaborative filtering: curse of dimension and long-consuming training. In our proposed algorithm, the curse of dimension problem is resolved by the proposed reduced-SVD technique effectively and long-consuming training is addressed by extreme learning machine (ELM) which is hundreds of times faster than iterative algorithms (e.g. BP). This will enable the algorithm more accurate and faster.