Assunção, João BorgesFernandes, Pedro Afonso2022-07-212022-07-212022-02-21http://hdl.handle.net/10400.14/38332This working paper proposes a new, practical method to compute the non-linear Mosheiov-Raveh (MR) filter using least absolute deviations (LAD) instead of the linear programming approach proposed by these two authors. This paper is embodied with an implementation in the R programming language of the proposed method which facilitates the computation of the MR filter in current applications to produce a robust estimate, namely, of the GDP trend growth. This technique may be appropriate to deal with non linear time series or structural changes.engBusiness cyclesNon linear time seriesRobust filteringSoftwareworking paper