📈 Naïve Baseline Forecast

Naïve Baseline Forecast#

A baseline predicting the next value equals the most recent observation.

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 Despite careful review, it may still contain mistakes, omissions, or out-of-date information.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#

A naïve baseline forecast is the simplest possible forecast — use the last observed value as the prediction for the next period (the random-walk forecast). Its seasonal cousin, the seasonal naïve, uses the value from one season ago:

\[\hat{y}_{t+1} = y_t \qquad\text{(naïve)}, \qquad\qquad \hat{y}_{t+h} = y_{t+h-m} \quad\text{(seasonal naïve, period } m).\]

Why it matters#

It is the benchmark every model must beat. If a complex model cannot outperform “just repeat the last value”, the complexity is not paying off. It is also the reference in the denominator of MASE, computed on the in-sample (training) series so the benchmark does not leak future information.

Which variant#

The plain naïve suits non-seasonal data; the seasonal naïve is the right reference when there is a clear period (set \(m = 12\) for monthly data with yearly seasonality, not \(m = 1\)). Cheap, robust, and — for noisy or short series — surprisingly hard to beat.

Theme: Signal Processing & Time Series · All terminology

Hint

Mind map — connected ideas

Forecast Error · Simple Baseline Methods · Seasonal Lag · Forecasting Benchmarks · M-Competitions (Makridakis Competitions) · Point Forecasts

Hint

More in Signal Processing & Time Series

ARIMA (AutoRegressive Integrated Moving Average) · Bayesian Time Series · Forecast Error · Forecasting Benchmarks · Forecasting Competitions · Log-Space · Low-pass Filtering · LSTM — Long Short-Term Memory Networks · M-Competitions (Makridakis Competitions) · Prophet — Time Series Forecasting by Facebook (Meta) · Seasonal Lag · Seasonality · Signal Processing · Simple Baseline Methods