Open eminemichael opened 5 years ago
When P=1, how is feature vector of this user trained, won't it be unchanged?
I don't think so. Since V changes, U will change.
我想的是P=1,在取k-1份作为训练集后,验证集中所有用户的特征向量在损失函数就不出现了,只有训练集中的用户向量,这样怎么对验证集的用户向量更新呢
When P=1, how is feature vector of this user trained, won't it be unchanged?