• Contents:

Feature Engineering - Basic Concepts

Ref: https://www.udemy.com/course/aws-certified-machine-learning-engineer-associate-mla-c01/learn/lecture/45284887

• 🔧  Applying domain knowledge (your knowledge of the data – and the model you’re using) to create better features to train your model
  • ‼️ ART OF ML!!
    • Most critical part in a good ML implementation
    • Talented/expert ML specialists are good at feature engineering
• Curse of dimensionality
  • ❗More features ≠ better!
    • Every feature is a new dimension
    • Much of feature engineering is selecting most relevant features → domain knowledge comes into play
  • Unsupervised dimensionality reduction techniques can help (PCA, K-Means)
• Common problems that Feature Engineering usually addresses:
  • Missing Data
  • Unbalanced Data
  • Outliers
  • Data Transformation

Missing Data

Ref: https://www.udemy.com/course/aws-certified-machine-learning-engineer-associate-mla-c01/learn/lecture/45285025

• 💡 Impute missing data = fill missing data with something

Impute: Mean Replacement

• 🔧 Replace missing values with mean value of column
  • 💡 A column represents a single feature
  • Median value of column can be more useful if outliers distort the mean
    • 💡 e.g. outlier billionaires distorting the income data of average citizens
• 👍 Pros
  • Fast & easy
  • Doesn't affect mean or sample size of overall data set
• 👎 Cons: pretty terrible
  • Not very accurate
  • Misses correlations between features (only works on column level)
    • If age & income are correlated, simply imputing the mean will muddy that relation a lot
  • Mean/median can only be calculated on numeric features, not on categorical features
    • 💡 Most frequent value in a categorical feature could work though
• Example code

Dropping Missing Data

• Reasonable if (all must apply!):
  • Not many rows with missing data
  • Dropping those rows doesn't bias data
  • Need a fast solution
• Almost anything is better though, rarely “best” approach
  • e.g. impute similar field (impute “review summary” into “full text”)
  • Data is generally valuable, dropping it is generally a bad idea