Loosely speaking, feature hashing uses a random sparse projection matrix A : Rn → Rm (where m ≪ n) in order to reduce the dimension of the data.
確定! 回上一頁