Applied Machine Learning: Feature Engineering

Extract the maximum value from your data using feature engineering. Learn how to clean, normalize, and create features to improve the performance of your machine learning models.

Introduction

• Applied ML: Feature engineering
• What you should know

1. Basic Techniques

• Imputation
• Filling in missing values
• Binning
• Log transform
• Scaling
• Challenge: Basic techniques
• Solution: Basic techniques

2. Categorical Encoding

• One hot encoding
• Hashing encoder
• Mean target encoding
• Challenge: Categorical
• Solution: Categorical

3. Feature Extraction

• PCA
• Feature aggregation
• TFIDF
• Text embeddings
• Challenge: Feature extraction
• Solution: Feature extraction

4. Temporal Features

• Extracting date components
• Seasonality and trend decomposition
• Challenge: Temporal features
• Solution: Temporal features

5. Feature Evaluation

• Importance and weights
• Recursive feature elimination
• Adding a random column
• Challenge: Feature selection
• Solution: Feature selection

Conclusion

• Next steps