[Python] Using pd.from_dummies to get the values in target variable back (after one-hot...

I'm participating in the curent Kaggle playground series and I decided to use a neural network for the binary classification problem because I am practicing making simple ones. I converted the target variable 'Personality' into 1s and 0s using OHE.

train2 = pd.get_dummies(train, columns=['Personality'], drop_first=True, dtype=int)


Then I used the tensorflow-keras library. The output was in decimals (because it's probability) and so I rounded up using:

pred = np.round(pred)



then converted the result to int64 and a dataframe with this code:

df = pd.DataFrame(pred)
pred_df = df.astype('int64')
pred_df.head()


but when I tried to revert the predicted values back to Introvert and Extrovert,

final = pd.from_dummies(pred_df)


I got an error.

This is the full error message ->
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
/tmp/ipykernel_635/341381991.py in <cell line: 0>()
1 #Revert
----> 2 final = pd.from_dummies(pred_df)

/usr/local/lib/python3.11/dist-packages/pandas/core/reshape/encoding.py in from_dummies(data, sep, default_category)
550 cats.append(default_category[prefix])
551 else:
--> 552 raise ValueError(
553 "Dummy DataFrame contains unassigned value(s); "
554 f"First instance in row: {assigned.idxmin()}"

ValueError: Dummy DataFrame contains unassigned value(s); First instance in row: 0


I'm not sure how to solve this or where I went wrong so that's why I put in as much code as I could. If reverting is impossible and just using a map function then an 'if 1 then Introvert' statement at the end is more plausible then I am willing to do that.

Thank you.

Continue reading...