search

mljar / mljar-supervised

Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation
https://mljar.com
MIT License
3.05k stars 406 forks source link

Add support for time/date features #122

Closed pplonski closed 4 years ago

pplonski commented 4 years ago

Related to https://github.com/mljar/mljar-supervised/issues/23

pplonski commented 4 years ago

mljar-supervised supports now date/time features! :tada: Feature should be translated to Pandas datetime before providing it to AutoML:

df["date"] = pd.to_datetime(df["date"])

Then AutoML will detect datetime features and will transform it by adding numerical features like: year, month, day, weekday, dayofyear, hour, days_diff_to_min (the difference between minimum date treated as reference). The new features will be added if there will be more than 1 unique value. For example, the AutoML will try to construct a year feature, but if all rows will have the same year then such feature will not be added.