Hi, would the organizer please provided a description on how the data is preprocessed in the provided code? Thanks~Posted by: harryfootball @ Aug. 10, 2018, 10:33 a.m.
Sure, as you can see in the starting kit (DataManager.py), time features are loaded as integers, numerical features as float64 variables, categorical and multi-valued categorical features are loaded using an ordinal encoding. No other preprocessing is made.
The return value for ordinal encoding for categorical features is one ordered integer feature
What is the return value of ordinal encoding for multi-valued categorical features?