Skip to content
Predict.ai

Fundamentals

Trend

Trend is the slow drift in level — up, down, or flattening — after you look past the weekly wiggle. Growth, decay, saturation. It is the part of the series that would still be there if you blurred out the season.

A new app's daily actives can rise for eighteen months and then bend. A factory's scrap rate can creep down after a process change. If you fit only last week's pattern, you will miss the bend until it is expensive.

Trends end. Linear-forever is a comfortable lie. When a series is saturating, a model that loved last year's slope will overshoot. Watch for a changing slope, not only a pretty line.

De-trending is optional. Some models want a stationary leftover. Others (and most foundation models) will eat trend as-is. You still want to see it with your eyes, because a trend you did not notice will be credited to a promo you did.

One series, four layers

Time series decomposed into trend, season, and leftoverObservedTrendSeasonLeftover
Most business series are a slow drift, a repeating clock, and leftover noise. Models differ in how explicitly they split these. You still want to see the split.

Read next

Related terms

Try it on your data.

Connect an outcome and its history. Predict.ai finds the drivers, races the models, and keeps the forecast live — in the workspace, over the API, and through agents.