📈  Log-Space

Log-Space#

Working with log-transformed values to stabilise variance or handle scale.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

Working in log-space means transforming a series to its logarithm, \(w_t = \log(y_t)\), instead of the raw values — a variance-stabilizing transform for when fluctuations grow with the level of the series (multiplicative or heteroscedastic behavior).

What it does#

It compresses large values while leaving small ones nearly untouched, turns multiplicative structure into additive, and makes relative (percentage) change the natural unit — a difference in log-space is approximately a proportional change:

\[\log(y_t) - \log(y_{t-1}) \approx \frac{y_t - y_{t-1}}{y_{t-1}}.\]

Caveats#

The logarithm is undefined for zero or negative values — use \(\log(y + 1)\) or a Box-Cox transform instead — and any forecast made in log-space must be back-transformed (exponentiated) to the original scale.


Theme: Signal Processing & Time Series  ·  All terminology



See also

Source article Adapted (context, re-expressed) in our own words from: Log-Space (insightful-data-lab.com).

Tags: purpose: reference topic: terminology level: advanced