论文部分内容阅读
One way of generation smooth approximations using scattered data in high dimension is by the moving least-squares method.We suggest to use an error measure which is between the ℓ1 and the ℓ2 norms,with the advantages of both.