Moving Averages & Smoothing in AI & Artificial Intelligence

A Simple Moving Average (SMA) is the most intuitive smoothing technique. It takes the mean of the values within a fixed 'window' of time (e.g., the last 7 days). This effectively 'filters out' short-term volatility. The larger the window, the smoother the line, but also the greater the Lag—the delay between a real-world change and that change appearing in the moving average.

Unlike the SMA, which treats all days in the window equally, the Exponential Moving Average (EMA) applies weights that decrease exponentially over time. This means the most recent data points have the most influence on the average. EMAs are widely used in financial trading because they react more quickly to price reversals while still providing a smoother signal than raw data.

Selecting a window size is a delicate balance. A Small Window (e.g., 3-5 periods) is highly responsive to new information but may capture too much noise. A Large Window (e.g., 50-200 periods) provides a very stable trend but can be significantly lagged, potentially missing a trend reversal until it's too late. The 'correct' window depends entirely on the frequency of your data and the goals of your analysis.

Sometimes the signal is buried in too much noise. 'Smoothing' allows us to blur out the random spikes and see the true underlying trend. The most popular tool for this is the Moving Average.

A Simple Moving Average (SMA) calculates the average of the last 'n' periods. As the window moves forward, the oldest data point is dropped and the newest is added.

While SMA treats all points equally, an Exponential Moving Average (EMA) gives more weight to recent data. It's more responsive to sudden changes in the trend.

Checkpoint: Which type of moving average reacts more quickly to a sudden price spike?

• →Simple Moving Average (SMA)
• →Exponential Moving Average (EMA)

Choosing the right 'window' is a trade-off. A small window keeps your data responsive but noisy; a large window creates a smooth line but 'lags' behind the actual data.

Smoothing isn't just for visualization. It's a key preprocessing step that helps your forecasting models focus on the patterns that matter.

Checkpoint: What happens to the smoothness of a line as you increase the 'rolling window' size?

• →It becomes rougher and noisier
• →It becomes smoother but more lagged

Smoothing mastered! You've learned to filter the noise. Ready to build predictive features for your time-series models?

Compute a Real Moving Average. Finish computing the simple moving average across a sliding 3-point window.