[FIXED] How to know expected input and output in pickle file from scikit-learn

Issue

Here's how I load pickle file

f = open('latest_model_v1_2.pkl', 'rb')
mdl = pickle.load(f)

and the mdl is

logisticRegressionCV(Cs=[0.09090909090909091, 0.18181818181818182,
                         0.2727272727272727, 0.36363636363636365,
                         0.45454545454545453, 0.5454545454545454,
                         0.6363636363636364, 0.7272727272727273,
                         0.8181818181818182, 0.9090909090909091],
                     cv=5, max_iter=10000, penalty='l1', solver='liblinear')

I don't know the expected name of variable input thos machine learning model, how to do this?

Solution

To know what parameters are expected in a parquet file, you can use feature_names_in_ variable.

pickled_model = pickle.load(open("path/to/pickle/file.pickle", "rb"))
pickled_model.feature_names_in_

output:

array(['engines', 'passenger_capacity', 'crew', 'review_scores_rating'],
      dtype=object)

Answered By - afzalex