Table of content

1. What is feature engineering ?
2. What happens when feature engineering is done ?
3. Types of features/variables.
   • 3.1 Numerical feature
   • 3.2 Categorical feature
   • 3.3 Date time feature
   • 3.4 Mixed feature
4. Feature Characteristics
   • 4.1 Missing values
   • 4.2 Cardinality
   • 4.3 Rare Labels
   • 4.4 Linear model assumptions
   • 4.5 Outliers
   • 4.6 Variable magnitude
5. Missing value imputation
   • 5.1 Complete case analysis
   • 5.2 Drop missing value observations
   • 5.3 Mean & median imputation
   • 5.4 Arbitary value imputation
   • 5.5 End of distribution imputation
   • 5.6 Frequent or mode category imputation
   • 5.7 Random sample imputation
   • 5.8 Missing indicator
6. Categorical encoding
   • 6.1 One hot encoding
   • 6.2 One hot encoding of frequent categories
   • 6.3 Integer encoding
   • 6.4 Count or frequency encoding
   • 6.5 Target ordered mean integer encoding
   • 6.6 Target mean encoding
   • 6.7 Target probability ratio encoding
   • 6.8 Target weight of evidence encoding
   • 6.9 Engineering rare categories
7. Feature/Variable Transformation
   • 7.1 Gaussian transformation
8. Discretisation
   • 8.1 Equal width discretisation
   • 8.2 Equal frequency discretisation
   • 8.2 K-means discretisation
   • 8.3 Discretisation plus encoding
   • 8.4 Discretisation using decision tree
   • 8.5 Discretisation based on domain knowledge
9. Outlier engineering
   • 9.1 Trim outlier
   • 9.2 Capping & IQR proximity rule
   • 9.3 Capping gausian approximation
   • 9.4 Capping quantiles
   • 9.5 Capping arbitrary
10. Feature scaling
    • 10.1 Standardisation
    • 10.2 Mean normalisation
    • 10.3 MinMaxScaling
    • 10.4 Maximum absolute scaling
    • 10.5 Robust scaling
    • 10.6 Scaling to unit length
11. Feature engineering mixed variable
    • 11.1 Engineering mixed variable
12. Feature engineering date and time
    • 12.1 Engineering date
    • 12.1 Engineering time
13. References & credits
14. Connect with me

---

Note 📜

• Feature and Variable are used interchangeably. Feature and Variable convey same meaning.
• Please find what, why & how part of specific topic inside the file "filename.py".

1. What is feature engineering 👨‍💻 ❓

• One reason for success of a ML project is coming up with a good set of features to train on.
• The process of obtaining a good set of features is called as feature engineering.

2. What happens when feature engineering is done ❓

• Improve performance of machine learning model
• Feature engineering represents data

13. References & Credits 🤝

• Feature Engine
• Feature Engineering for Machine Learning

14. Connect with me 😃