Performance-optimal filtering

Optimal filtering

This paper presents short summary about commonly used filters in macroeconomics. We consider Hodrick-Prescott filter, Butterworth filter, Ideal Bandpass filter, Baxter-King filter and Christiano-Fitzgerald filter. 1 The filtering problem Given raw data, {xt}t=1, we are interested in isolating component of xt, denoted by {yt}t=1, with period of oscillation between pl and pu, where 2 ≤ pl < pu < ...

Optimal cumulant domain filtering

This paper presents a new technique which exploits constrained optimization methods to derive optimal two dimensional filters in the cumulant domain for processing signals in non Gaussian noise, or signals with corrupting interferences which have non symmetrical probability density functions. The approach proposed here for enhancing signals in such noise is important, as increasingly practical ...

Connections between H2 optimal filters and unknown input observers – Performance limitations of H2 optimal filtering

This paper establishes a connection between the unknown input observer problem (also known as the exact input-decoupling filtering problem) and the Kalman filtering problem. Such a connection leads to a better understanding of filtering. As a consequence of this, tools for the analysis as well as synthesis of filters can be developed. Moreover, such tools can be utilized to establish performanc...

Optimal Seasonal Filtering

In this paper, I compare model-based and moving average-based approaches to seasonal adjustment. I propose an optimal moving average filter, designed to minimize the distance from a model-based filter. In simulations, I consider the accuracy of the alternative methods in estimating seasonal factors in a simple model. I find that the most reliable results are obtained either by the model-based a...

H2 Optimal Robust Filtering