I’m not sure what is the context of this question. Are you asking about the technical meaning of filtration in machine learning?
The most common meaning of filtration in machine learning is in “time series modelling”, which are machine learning systems which operate with data that is coming in a stream, e.g. financial time series of stock prices over time. A filter is a system that extracts knowledge from past data to predict into the future. A popular system is called a Kalman filter in this context.
Comments