MMD_4c_LatentFactor

deal more generally with the problem of decomposing very large matrices into products of matrices with at least one small dimension. The small dimension corresponds to the hidden factors like movie genre.

• more on svd
  • how to query
  • relation to eigen-decomposition
  • drabacks
• CUR
• Latent factor model
  • netflix prize
  • local and global effects
  • rmse of various methods
• Latent factor RS
  • model
  • differ with svd
  • finding latent factor

more on svd

how to query

这里写图片描述

relation to eigen-decomposition

这里写图片描述

drabacks

这里写图片描述

CUR

Latent factor model

netflix prize

这里写图片描述

local and global effects

这里写图片描述

rmse of various methods

这里写图片描述

Latent factor RS

model

这里写图片描述

这里写图片描述

differ with svd

svd需要矩阵A的元素都有，不能有missing value，但是latent factor analysis没有这些要求。

这里写图片描述

finding latent factor

这里写图片描述