求帮助：对非参数模型y_i=f(x_i)+e_i,用leave-one-out估计量hat_(f_{-i}(x_i))估计f, 此时用cross validation方法选择窗宽h时，即
minimize(y_i-hat_(f_{-i}(x_i)))^2,怎样用R语言编程？谢谢