Data for machine learning · Question 37 of 60
Which tasks belong to data preprocessing?
3 answers are correct. Medium
Show answer
Answer
- A Cleaning data, for example removing duplicates and outliers
- B Sampling subsets to reduce training times and computational costs
- C Augmenting data, for example with adversarial examples or synthetic data
Explanation
Preprocessing includes cleaning, transforming formats, scaling and normalizing, augmenting (more samples, adversarial examples for robustness, synthetic data) and sampling. Capturing data belongs to data acquisition, and extracting features to feature engineering.