#train-test-split

Question 1: Select all that apply

Which of these setups leak information from the evaluation data into training, so that the measured test performance is too optimistic? Select all that apply.

Standardizing all features with the mean and standard deviation of the full dataset, then splitting into training and test sets1
Splitting product photos randomly by image, when each product has several photos, so the same product appears in training and test2
Fitting the scaler on the training split only and applying those statistics to the validation and test splits3
Forecasting demand by training on January–September and testing on October–December4

Timed test