4.2.3. Data Preprocessing and Feature Engineering

Try these first

1. 1.

   What is data cleaning?

   Hint

   Think about why clean data is critical.

2. 2.

   Why do we normalize data?

   Hint

   Consider how different units could affect learning.

3. 3.

   What is the primary purpose of data cleaning?

   • To improve model accuracy
   • To fix inconsistencies in data
   • To generate new features

   Hint

   Consider the initial step before any model training.

4. 4.

   True or False: Feature engineering is unnecessary if the dataset is large.

   • True
   • False

   Hint

   Think about how features impact model performance.

5. 5.

   Given a dataset with numerous missing values and outliers, outline a detailed plan to preprocess this data for training a machine learning model.

   Hint

   Break it down into cleaning, engineering features, and then scaling.

6. 6.

   How can poor feature engineering lead to the failure of an AI application? Provide an example.

   Hint

   Consider how representation of data informs learning.