Lasso trigonometric polynomial approximation for periodic function recovery in equidistant points (Q6086846)

The authors study in this paper Lasso Trigonometric interpolation (LTI). They study approximation properties of LTI and provide error analysis. They show that in the absence of noise, the L2 error of LTI for any nonzero periodic continuous function converges to a nonzero term which depends on the best approximation of this function, whereas such an error of trigonometric interpolation on an equidistant grid converges to zero as $N\to\infty$. But in the presence of noise, LTI is able to reduce the newly introduced error term caused by noise by multiplying a factor less than one.