Do you want to use Exponential Smoothing but your data has a trend?

Double Exponential Smoothing is here to save the day.

🧵 THREAD🧵

👇👇👇👇

#MachineLearning #DataScience #Python #TimeSeries #AI #ML

This extension of the Simple Exponential Smoothing model adds support for modelling the trend, making it suitable for data with a trend but no seasonality.

There are two kinds of models depending on the type of trend:

🔹additive trend (Holt's linear trend model)

🔹multiplicative trend

The additive trend model uses Double Exponential Smoothing with a linear trend.

While the multiplicative trend model uses Double Exponential Smoothing with an exponential trend.

Both models have a set of equations that include variables such as level value, trend value, growth rate value, and smoothing constants for the trend.

We can inspect the additive model equations:

Thanks to the statsmodels library in Python, implementing this model is a breeze.

Just fit your data, set your smoothing constants, and forecast the next few periods.

This model is capable of forecasting both the level and the trend, so you can see how the forecasting has a slope.

So if you're tired of boring forecasting models that can't handle trends, give Double Exponential Smoothing a try!

Check more about Exponential Smoothing in this week's article:

mlpills.dev/time-series/exponential-smoothing/

You can subscribe to our 📩 Newsletters so you don't miss out on any of our articles and additional content!

🚀DSBoost:
dsboost.substack.com/

💊MLPills:
mlpills.substack.com/