They call it by a different name (of course, because they're economists), but it's the same thing.
towardsdatascience.com
This is the same thing that goes on in the hippocampus of the brain, where times and locations are encoded.
Yes, it's a valuable method for machine learning, it enables very efficient data storage in neural networks, and also makes the information available in conditions of criticality (where it plays like the Kuramoto model of coupled oscillators)
Cyclical Encoding: An Alternative to One-Hot Encoding for Time Series Features
Cyclical encoding provides your model with the same information using significantly fewer features
towardsdatascience.com
This is the same thing that goes on in the hippocampus of the brain, where times and locations are encoded.
Yes, it's a valuable method for machine learning, it enables very efficient data storage in neural networks, and also makes the information available in conditions of criticality (where it plays like the Kuramoto model of coupled oscillators)
