Linear Predictive Filters for Dominant Cycle Estimation

Summary

This indicator implements linear predictive filter methods for examining cycles in market time series. A second-order high-pass filter and a Super Smoother band-limit the source data, after which an adaptive predictor updates coefficients from recent signal errors. The script offers a forward prediction, a spectrum across a configurable period range, and a dominant-cycle estimate. The spectrum evaluates predictor coefficients at candidate frequencies and normalizes the resulting power values for display; the dominant-cycle output selects the strongest candidate and limits abrupt changes from one bar to the next.

The tool includes an optional synthetic cycle signal for testing and lets users choose which outputs to display. Its method follows a signal-processing approach intended to identify changing cyclic structure, but the provided material does not demonstrate trading profitability or compare the method with alternatives. Cycle estimates depend on filter bounds and model length, and market prices may not contain stable or exploitable cycles, so the outputs should be treated as analytical inputs rather than standalone trade signals.

Key ideas

• High-pass and smoothing filters restrict analysis to a selected range of cycle lengths.
• An adaptive linear predictor updates coefficients using recent forecast errors.
• The indicator displays a forward signal prediction, a frequency spectrum, and a dominant-cycle estimate.
• The dominant-cycle output limits abrupt changes between successive estimates.
• Cycle detection is parameter-dependent and the document gives no evidence of trading performance.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.