KEMD: Kipnis Empirical Mode Decomposition

In IlyaKipnis/DSTrading: Digital Signal Trading

Description

Kipnis Empirical Mode Decomposition

Usage

  KEMD(x, delta = 0.1, n = 20, bandFraction = 0.1,
    maType = "SMA", ...)

Arguments

x	a time series
delta	bandpass parameter (see reference paper), default .1
n	a lookback period, default 20
bandFraction	trending vs. mean-reverting band control. Raise to bias towards mean reversion. Lower to bias towards trend-following. Defaults to .1
maType	type of moving average to use. Defaults to SMA.

Value

an xts consisting of a trend, a peak band, a valley band, a pctB, and momentum. The market is classified as trending upwards when the trend is above the peak band, trending downwards when below the valley, and rangebound in between. The pctB is a different way of looking at the relationship between the trend and the peak and valleys. The momentum column indicates trend direction.

Note

– I (Ilya Kipnis) made a special modification compared to Ehlers that divides the peak and valley bands by an running n-point standard deviation of the bandpass. This helps make the bands more adaptive and less likely to give whipsaw signals compared to Ehlers's Empirical Mode Decomposition. PctB and momentum are added features as well.

References

http://www.mesasoftware.com/Seminars/Trend%20Modes%20and%20Cycle%20Modes.pdf

IlyaKipnis/DSTrading documentation built on May 8, 2019, 1:39 p.m.