Implementation:Online ml River Stats PeakToPeak

Knowledge Sources	Online_ml_River
Domains	Online_Learning, Statistics
Last Updated	2026-02-08 16:00 GMT

Overview

PeakToPeak computes the running peak-to-peak range (maximum minus minimum) of a data stream.

Description

This statistic calculates the difference between the maximum and minimum values observed in streaming data. It provides a simple measure of the total range or spread of the data. The implementation includes PeakToPeak for the overall range and RollingPeakToPeak for the range within a sliding window. The basic version is optimized using Rust for performance.

Usage

Use PeakToPeak when you need a quick measure of data spread or range. Common applications include monitoring signal amplitude, detecting changes in data variability, quality control for checking if values stay within expected bounds, and normalization where you need to know the full range of values.

Code Reference

Source Location

• Repository: Online_ml_River
• File: river/stats/ptp.py

Signature

class PeakToPeak(stats.base.Univariate):
    def __init__(self):
        self._ptp = _rust_stats.RsPeakToPeak()
        self._is_updated = False

class RollingPeakToPeak(stats.base.RollingUnivariate):
    def __init__(self, window_size: int):
        self.max = stats.RollingMax(window_size)
        self.min = stats.RollingMin(window_size)

Import

from river import stats

I/O Contract

Inputs

Name	Type	Required	Description
x	numbers.Number	Yes	Value to update the statistic with
window_size	int	Yes (for Rolling)	Size of the rolling window

Outputs

Name	Type	Description
get()	float	Current peak-to-peak range (max - min)

Usage Examples

from river import stats

# Basic peak-to-peak
X = [1, -4, 3, -2, 2, 4]
ptp = stats.PeakToPeak()

for x in X:
    ptp.update(x)
    print(f"Value: {x}, Peak-to-Peak: {ptp.get():.1f}")

# Output:
# Value: 1, Peak-to-Peak: 0.0
# Value: -4, Peak-to-Peak: 5.0
# Value: 3, Peak-to-Peak: 7.0
# Value: -2, Peak-to-Peak: 7.0
# Value: 2, Peak-to-Peak: 7.0
# Value: 4, Peak-to-Peak: 8.0

# Rolling peak-to-peak
X = [1, -4, 3, -2, 2, 1]
rolling_ptp = stats.RollingPeakToPeak(window_size=2)

for x in X:
    rolling_ptp.update(x)
    print(f"Value: {x}, Rolling P2P: {rolling_ptp.get()}")

# Output:
# Value: 1, Rolling P2P: 0
# Value: -4, Rolling P2P: 5
# Value: 3, Rolling P2P: 7
# Value: -2, Rolling P2P: 5
# Value: 2, Rolling P2P: 4
# Value: 1, Rolling P2P: 1

# Using for signal amplitude monitoring
amplitude_monitor = stats.PeakToPeak()
signal = [0, 5, 10, 5, 0, -5, -10, -5, 0]

for s in signal:
    amplitude_monitor.update(s)

print(f"Signal amplitude: {amplitude_monitor.get():.1f}")
# Output: 20.0 (from -10 to 10)

# Range-based normalization
ptp_norm = stats.PeakToPeak()
min_val = stats.Min()

data = [10, 25, 15, 30, 5, 20]
for x in data:
    ptp_norm.update(x)
    min_val.update(x)

# Normalize to [0, 1]
range_val = ptp_norm.get()
min_v = min_val.get()
normalized = [(x - min_v) / range_val for x in data]
print(f"Normalized: {[f'{n:.2f}' for n in normalized]}")

Related Pages

• Environment:Online_ml_River_Python_Runtime_Environment