Open Yuhei-Handa opened 6 months ago
Thank you for your message! 😊
It's great to see that you've addressed the data leakage issue in my preprocessing section. By splitting the data into training and testing sets before applying the Min-Max scaling, I've ensured that the scaling parameters are learned only from the training data and then consistently applied to the test data. This approach helps prevent any information leakage from the test set into the training process.
Here's a brief description of the changes I made:
data_df
) to a NumPy array (data_numpy
).x_train
, y_train
) and testing (x_test
, y_test
) sets using train_test_split
.x_train_scaled
) using a feature range of (-1, 1).x_test_scaled
) using the same scaler.
I have a question about the preprocessing section in the ipynb.
It seems that the preprocessing of min-max scaling is applied to the entire dataset, but it is not distinguished between training data and validation data, which may cause data leakage. How are you addressing this issue?