Learn More
This notes cover dimensionality reduction of the linear regression problem using subspace embedding, and background for the streaming JL problem. The linear regression problem is defined as follows: Given a matrix A ∈ R n×d and a vector b ∈ R n , find x such that: x = argmin x∈R d Ax − b 2. In this case, a closed formula is known for x : x = (A T A) −1 A T(More)
  • 1