For a properly scaled continuous positive-definite shift-invariant kernel, drawing frequencies from the kernel's Fourier transform produces random sinusoidal features whose inner product is an unbiased estimator of the kernel; averaging D independent frequency draws lowers estimator variance. · CiteArk
For a properly scaled continuous positive-definite shift-invariant kernel, drawing frequencies from the kernel's Fourier transform produces random sinusoidal features whose inner product is an unbiased estimator of the kernel; averaging D independent frequency draws lowers estimator variance.
来源：paper:PDF pp. 2-4, Section 3, Eq. (2), Figure 1, and Algorithm 1
