I have a dataframe which the target features, which looks like this:
x x1 y
1 2 3
2 3 4
Now I use feautretools to automatically do feature engineering, using this line of code:
es = ft.EntitySet(id = 'x')
es.entity_from_dataframe(entity_id = 'y', dataframe = df, index = 'x')
feature_matrix, feature_names = ft.dfs(entityset=es,
target_entity = 'y',
max_depth = 2,
verbose = 1,
n_jobs = 3)
I would like to take the features generated, and then apply them to a dataset which lacks the labels, something which looks like this:
x x1
1 2
How would I take the features generate (e.g mean of x + x1) and then map their creation process ((df['x']+df['x1']).mean()) onto the dataframe lacking the label?
This answered my question, the saving feature part: https://featuretools.alteryx.com/en/stable/guides/deployment.html