Forecast

Method Library

All 12 Methods

Click any method to read the full page — when to use, when not to use, Python code, and a real finance example.

01

Naive / Seasonal Naive

Benchmark; every other method must beat these

Difficulty · Trivial
02

SES (Simple Exponential Smoothing)

Flat, noisy series with no pattern

Difficulty · Easy
03

Holt (Linear Trend)

Trending series, no seasonality

Difficulty · Easy
04

Holt-Winters

Trending + seasonal series; wins more often than expected

Difficulty · Easy
05

Theta

Strong simple baseline; top-10 in M3/M4; underused

Difficulty · Easy
06

ARIMA / Auto-ARIMA

Flexible univariate; linear trend

Difficulty · Medium
07

SARIMA

Univariate with seasonality + trend

Difficulty · Medium
08

TBATS

Multiple/complex seasonal cycles

Difficulty · Medium
09

Prophet

Fast deployment with messy data; verify against simpler methods

Difficulty · Easy
10

LightGBM with features

M5 winning approach; many series with shared structure

Difficulty · Medium-Hard
11

Croston / TSB / iETS

Intermittent demand (many zeros)

Difficulty · Easy
12

Ensembling

Almost always beats picking one method; default for production

Difficulty · Medium

Back to the Matrix