This is the very well-known wavelet based denoising method (see reference (1)): The essential idea is that "meaningful" signals have their energy concentrated in few significant wavelet coefficients, while white noise, to the contrary, has coefficients which all behave the same in the statistical sense. The denoising is then performed by fixing a threshold and setting to 0 all coefficients which are below this threshold. Coefficients above it are shrinked towards 0, i.e. the value of the threshold is subtracted to them (if they are positive, with obvious modification for negative coefficients). With a right choice of the threshold, this procedure has been proved to be asymptotically rate-minimax for additive noise.
Again, specify your Analyzed signal, and the desired Threshold Factor. Then hit Compute to get the regularized signal, called den_signal#. Note that typical values for the threshold are around 0.1 or below.