I'm trying to use ML algorithm to do classification on time series data and streaming data. Although I'm able to find certain ML algorithms applicable to such data, such as dynamic time warping, I think they will achieve a better accuracy if feature engineering is performed, and my goal is to use feature engineering to convert the time series data into IID feature sets then use the common machine learning libraries such as random forecast to do the work on these IID feature sets.
Therefore, my goal is: build as many summarizers of the time series data as possible, such as median/mean/max and their rolling window correspondance, then use these features as input to machine learning libraries for IID data such as random forecast.
I am not able to find any comprehensive introduction on the feature engineering techniques for time series data and streaming data like this. Can someone share with me some common techniques, such as common transformations on such kind of data?
