Moving average features

In statistics, a moving average (rolling average or running average) is a calculation to analyze data points by creating a series of averages of different subsets of the full data set.

Available moving average features

Moving averages utilize past values for calculation of the current values, hence while it might be easy to create such features during training in a single go, productionalizing or implementation for test become a challenge.

NitroFE features overcome this challenge by internally saving the past dependant values (the number of to be saved depend upon the logic of the feature ). For your training/initial fit phase use fit_first=True, and for any production/test implementation pass fit_first=False

• exponential moving average
• hull moving average
• kaufman adaptive moving average
• fractal adaptive moving average
• triple exponential moving average
• smoothed moving average