A Simple Nonlinear Filter for Economic Time Series Analysis

Y Wen,B Zeng
DOI: https://doi.org/10.1016/s0165-1765(99)00089-0
IF: 1.469
1999-01-01
Economics Letters
Abstract:In applied macroeconomics, especially business cycle studies, the most commonly used filters for removing time trend and noise are linear filters. In this paper we introduce a class of nonlinear filters that has been recently proven very powerful for removing time trend and noise in signal processing. Interesting applications to economics are carried out in an example of forecasting post-war US recessions.
What problem does this paper attempt to address?