Running 📈 Approximations using random features Toy implementation of the RFM proposed by Nelsen and Stuart