Feature Engineering

#machine-learning #interview
Feature Engineering

Transforming raw data into feature vector is called feature engineering

Ways to do Feature Engineering:

  1. One Hot Vector
  2. Word Embeddings
  3. Feature Hashing
  4. Crossed Feature
  5. Numeric Features:
    1. Standardization
    2. Normalization
  6. Data Cleaning
    1. Handling Imbalanced Dataset
    2. Handling Missing Data
    3. Handling Outliers
    4. Deduplication

Related Notes