fabien-vavrand

Refactoring of the auto machine learning module

NumericalEncoder currently encodes categories to integers, without considering the integer type. Downcasting integers to int8 or int16 when possible could sometimes significantly reduce memory and avoid some memory overhead. import...

Bool columns are treated as categories, but should be transformed to integer instead import pandas as pd from aikit.transformers import NumericalEncoder df = pd.DataFrame(data={'a': [True, True, False]}) df['a'].dtype x =...