- In the case of local linear regression, let \(b(x_0)^T = (1,x_0)\) and note that \(b(x_0)^T = b(x_0)^T (\boldsymbol{B}^T\boldsymbol{W}(x_0)\boldsymbol{B})^{-1}\boldsymbol{B}^T\boldsymbol{W}(x_0)\boldsymbol{B} \). Then, re-express the right-hand side in terms of \(l(x_0)^T = (l_1(x_0),\ldots , l_N(x_0))\).
- In the case of local polynomial regression, let \(b(x_0)^T = (1,x_0,x_0^2,\ldots ,x_0^k)\) and do the same thing as above.
Hints for problem 6.2
Published Sep. 18, 2025
- Last modified Sep. 18, 2025