What is the Ehlers Autocorrelation Periodogram?

Ehlers Autocorrelation Periodogram is a cycle-estimation tool that turns autocorrelation structure into a smoothed dominant-cycle line and a companion normalized-power reading. VectorTA exposes both outputs across single-run, streaming, batch, Python, and WASM workflows.

How do I calculate the Ehlers Autocorrelation Periodogram?

The Ehlers Autocorrelation Periodogram is calculated using specific mathematical formulas with parameters: min_period (default: 8), max_period (default: 48), avg_length (default: 3), enhance (default: true).

What are the best settings for Ehlers Autocorrelation Periodogram?

The optimal settings depend on your trading style and timeframe. Default settings are: min_period: 8, max_period: 48, avg_length: 3, enhance: true. Experiment with different values to find what works best for your strategy.

Ehlers Autocorrelation Periodogram

Parameters: min_period = 8 | max_period = 48 | avg_length = 3 | enhance = true

Overview

Ehlers Autocorrelation Periodogram is not a trend oscillator in the usual sense. It is a cycle detector. The implementation high-pass filters and smooths the input series, measures how strongly that filtered signal correlates with lagged copies of itself, and then converts those lag relationships into a periodogram-like power surface across the candidate cycle lengths between the chosen minimum and maximum periods.

VectorTA distills that internal spectrum into two outputs. The dominant-cycle line is the smoothed winning cycle estimate, expressed in bars. The normalized-power line is the power value at the selected dominant-cycle bucket, which acts like a confidence read on how concentrated the spectral energy is around that chosen period. The optional enhance mode sharpens strong peaks by cubing normalized power before the final weighting step, making the dominant-cycle estimate less influenced by weaker candidate periods.

Defaults: Ehlers Autocorrelation Periodogram uses `min_period = 8`, `max_period = 48`, `avg_length = 3`, `enhance = true`, and defaults candle input to `close`.

Implementation Examples

Compute the dominant cycle and normalized power from a close series or a candle source.

use vector_ta::indicators::ehlers_autocorrelation_periodogram::{
    ehlers_autocorrelation_periodogram,
    EhlersAutocorrelationPeriodogramInput,
    EhlersAutocorrelationPeriodogramParams,
};
use vector_ta::utilities::data_loader::{Candles, read_candles_from_csv};

let output = ehlers_autocorrelation_periodogram(
    &EhlersAutocorrelationPeriodogramInput::from_slice(
        &close,
        EhlersAutocorrelationPeriodogramParams {
            min_period: Some(8),
            max_period: Some(48),
            avg_length: Some(3),
            enhance: Some(true),
        },
    )
)?;

let candles: Candles = read_candles_from_csv("data/sample.csv")?;
let candle_output = ehlers_autocorrelation_periodogram(
    &EhlersAutocorrelationPeriodogramInput::with_default_candles(&candles)
)?;

println!("dominant cycle = {:?}", output.dominant_cycle.last());
println!("normalized power = {:?}", candle_output.normalized_power.last());

The builder controls the candidate cycle range, the autocorrelation averaging window, and whether spectral peaks are enhanced.

use vector_ta::indicators::ehlers_autocorrelation_periodogram::EhlersAutocorrelationPeriodogramBuilder;
use vector_ta::utilities::enums::Kernel;

let defaultish = EhlersAutocorrelationPeriodogramBuilder::new()
    .min_period(8)
    .max_period(48)
    .avg_length(3)
    .enhance(true)
    .kernel(Kernel::Auto)
    .apply_slice(&close)?;

let source_based = EhlersAutocorrelationPeriodogramBuilder::new()
    .min_period(10)
    .max_period(60)
    .avg_length(5)
    .enhance(false)
    .kernel(Kernel::Scalar)
    .apply(&candles, "hl2")?;

println!("{:?}", defaultish.dominant_cycle.last());
println!("{:?}", source_based.normalized_power.last());

Streaming mode updates the dominant-cycle estimate and current power concentration one value at a time.

use vector_ta::indicators::ehlers_autocorrelation_periodogram::{
    EhlersAutocorrelationPeriodogramBuilder,
    EhlersAutocorrelationPeriodogramParams,
    EhlersAutocorrelationPeriodogramStream,
};

let mut stream = EhlersAutocorrelationPeriodogramStream::try_new(
    EhlersAutocorrelationPeriodogramParams {
        min_period: Some(8),
        max_period: Some(48),
        avg_length: Some(3),
        enhance: Some(true),
    }
)?;

for value in live_values {
    if let Some((dominant_cycle, normalized_power)) = stream.update(value) {
        println!("{dominant_cycle}, {normalized_power}");
    }
}

println!("warmup = {}", stream.get_warmup_period());

let builder_stream = EhlersAutocorrelationPeriodogramBuilder::new()
    .min_period(10)
    .max_period(60)
    .avg_length(5)
    .enhance(false)
    .into_stream()?;

Batch mode sweeps candidate cycle ranges and averaging windows while keeping the output focused on dominant cycle and its selected power.

use vector_ta::indicators::ehlers_autocorrelation_periodogram::EhlersAutocorrelationPeriodogramBatchBuilder;
use vector_ta::utilities::enums::Kernel;

let batch = EhlersAutocorrelationPeriodogramBatchBuilder::new()
    .min_period_range((8, 12, 2))
    .max_period_range((36, 48, 12))
    .avg_length_range((3, 5, 2))
    .enhance(true)
    .kernel(Kernel::Auto)
    .apply_slice(&close)?;

println!("rows = {}, cols = {}", batch.rows, batch.cols);
println!("first combo = {:?}", batch.combos.first());
println!("first dominant row = {:?}", &batch.dominant_cycle[..batch.cols.min(5)]);

API Reference

Input Methods ▼

// From candles and a named source field
EhlersAutocorrelationPeriodogramInput::from_candles(
    &Candles,
    &str,
    EhlersAutocorrelationPeriodogramParams,
) -> EhlersAutocorrelationPeriodogramInput

// From a raw slice
EhlersAutocorrelationPeriodogramInput::from_slice(
    &[f64],
    EhlersAutocorrelationPeriodogramParams,
) -> EhlersAutocorrelationPeriodogramInput

// From candles with default parameters
EhlersAutocorrelationPeriodogramInput::with_default_candles(&Candles)
    -> EhlersAutocorrelationPeriodogramInput

Parameters Structure ▼

pub struct EhlersAutocorrelationPeriodogramParams {
    pub min_period: Option<usize>, // default 8
    pub max_period: Option<usize>, // default 48
    pub avg_length: Option<usize>, // default 3
    pub enhance: Option<bool>,     // default true
}

Output Structure ▼

pub struct EhlersAutocorrelationPeriodogramOutput {
    pub dominant_cycle: Vec<f64>,
    pub normalized_power: Vec<f64>,
}

Validation, Warmup & NaNs ▼

The input slice must be non-empty and contain at least one finite value.
min_period must be at least 3.
max_period must be greater than min_period, and when real data length is known it cannot exceed that length.
The longest contiguous finite run must be at least warmup_period + 1 bars, where warmup_period = max_period + corr_window - 1.
Streaming resets on non-finite input and returns None until warmup completes again.
The normalized-power output is the power at the selected dominant-cycle bucket, not the full spectrum.
When enhance is enabled, normalized power is cubed before the dominant-cycle weighting step, making strong peaks dominate more aggressively.

Builder, Streaming & Batch APIs ▼

// Builder
EhlersAutocorrelationPeriodogramBuilder::new()
    .min_period(usize)
    .max_period(usize)
    .avg_length(usize)
    .enhance(bool)
    .kernel(Kernel)
    .apply_slice(&[f64])

EhlersAutocorrelationPeriodogramBuilder::new()
    .apply(&Candles, &str)

EhlersAutocorrelationPeriodogramBuilder::new()
    .into_stream()

// Stream
EhlersAutocorrelationPeriodogramStream::try_new(EhlersAutocorrelationPeriodogramParams)
EhlersAutocorrelationPeriodogramStream::update(f64) -> Option<(f64, f64)>
EhlersAutocorrelationPeriodogramStream::get_warmup_period() -> usize

// Batch
EhlersAutocorrelationPeriodogramBatchBuilder::new()
    .min_period_range((usize, usize, usize))
    .max_period_range((usize, usize, usize))
    .avg_length_range((usize, usize, usize))
    .enhance(bool)
    .kernel(Kernel)
    .apply_slice(&[f64])

EhlersAutocorrelationPeriodogramBatchBuilder::new()
    .apply_candles(&Candles, &str)

Error Handling ▼

pub enum EhlersAutocorrelationPeriodogramError {
    EmptyInputData,
    AllValuesNaN,
    InvalidMinPeriod { min_period: usize },
    InvalidMaxPeriod { max_period: usize, data_len: usize },
    InvalidPeriodOrder { min_period: usize, max_period: usize },
    NotEnoughValidData { needed: usize, valid: usize },
    OutputLengthMismatch {
        expected: usize,
        dominant_cycle_got: usize,
        normalized_power_got: usize,
    },
    InvalidRange { start: String, end: String, step: String },
    InvalidKernelForBatch(Kernel),
}

Python Bindings

Python exposes a two-array single-run function, a streaming class, and a batch function. The single-run binding returns the dominant-cycle array together with the normalized-power array. Batch returns both matrices plus the tested min periods, max periods, averaging lengths, enhance flags, and the rows and cols shape.

import numpy as np
from vector_ta import (
    ehlers_autocorrelation_periodogram,
    ehlers_autocorrelation_periodogram_batch,
    EhlersAutocorrelationPeriodogramStream,
)

data = np.asarray(close_values, dtype=np.float64)

dominant_cycle, normalized_power = ehlers_autocorrelation_periodogram(
    data,
    min_period=8,
    max_period=48,
    avg_length=3,
    enhance=True,
    kernel="auto",
)

stream = EhlersAutocorrelationPeriodogramStream(
    min_period=8,
    max_period=48,
    avg_length=3,
    enhance=True,
)
print(stream.update(data[-1]))
print(stream.warmup_period)

batch = ehlers_autocorrelation_periodogram_batch(
    data,
    min_period_range=(8, 12, 2),
    max_period_range=(36, 48, 12),
    avg_length_range=(3, 5, 2),
    enhance=True,
    kernel="auto",
)

print(batch["min_periods"], batch["enhance_flags"], batch["rows"], batch["cols"])

JavaScript/WASM Bindings

The WASM layer exposes an object-returning single-run wrapper, pointer-oriented in-place helpers, and a batch wrapper that returns both outputs together with parameter metadata. The single-run helper returns named dominant_cycle and normalized_power arrays. The batch helper adds combos, parameter vectors, rows, and cols.

import init, {
  ehlers_autocorrelation_periodogram_js,
  ehlers_autocorrelation_periodogram_batch_js,
} from "/pkg/vector_ta.js";

await init();

const data = new Float64Array(closeValues);

const single = ehlers_autocorrelation_periodogram_js(data, 8, 48, 3, true);
console.log(single.dominant_cycle, single.normalized_power);

const batch = ehlers_autocorrelation_periodogram_batch_js(data, {
  min_period_range: [8, 12, 2],
  max_period_range: [36, 48, 12],
  avg_length_range: [3, 5, 2],
  enhance: true,
});

console.log(
  batch.dominant_cycle,
  batch.normalized_power,
  batch.min_periods,
  batch.enhance_flags,
  batch.rows,
  batch.cols
);

CUDA Bindings (Rust)

Additional details for the CUDA bindings can be found inside the VectorTA repository.

Performance Analysis

Comparison:

View:

Placeholder data (no recorded benchmarks for this indicator)

Across sizes, Rust CPU runs about 1.14× faster than Tulip C in this benchmark.

Loading chart...

AMD Ryzen 9 9950X (CPU) | NVIDIA RTX 4090 (GPU)

Related Indicators

High-Low Correlation

Technical analysis indicator

Correlation Cycle

Technical analysis indicator

Deviation

Technical analysis indicator

Ehlers Error-Correcting EMA (ECEMA)

Moving average indicator

Ehlers Instantaneous Trendline

Moving average indicator

Ehlers Kaufman Adaptive Moving Average

Moving average indicator