Finding More Hidden Gems in Holt-Winters

Welcome back to this three-part blog post series on Holt-Winters and why it’s still highly relevant today. To understand Part Two, I suggest reading Part One, in which we covered:

1. When to use Holt-Winters;
2. How Single Exponential Smoothing works;
3. A conceptual overview of optimization for Single Exponential Smoothing;
4. Extra: The proof for optimization of Residual Sum of Squares (RSS) for Linear Regression.

In this piece, Part Two, we’ll explore:

1. How Single Exponential Smoothing relates to Triple Exponential Smoothing/Holt-Winters;
2. How RSS relates to Root Mean Square Error (RMSE);
3. How RMSE is optimized for Holt-Winters using the Nelder-Mead method.

In Part Three, we’ll explore:

1. How you can use InfluxDB’s built-in Multiplicative Holt-Winters function to generate predictions on your time series data;
2. A list of learning resources.

How Single Exponential Smoothing relates to Triple Exponential Smoothing/Holt-Winters

Like SES, Holt-Winters, which can be employed as a powerful and efficient predictive maintenance technique, determines the forecasted value by calculating an exponentially weighted average — but it doesn’t stop there.