Intuitive Explanation of Exponential Moving Average

Understand the logic behind the fundamental algorithm used inside the gradient descent

Introduction

In time series analysis, there is often a need to understand the trend direction of a sequence by taking into account previous values. Approximation of the next values in a sequence can be performed in several ways, including the usage of simple baselines or the construction of advanced machine learning models.

An exponential (weighted) moving average is a robust trade-off between these two methods. Having a simple …

algorithm analysis approximation getting-started gradient gradient-descent inside logic machine learning moving next series simple statistics time series time-series-analysis trend usage values