Machine Learning Zoomcamp Update: Thursday, 16 September 2021
Date: 16 September 2021
Today, I completed the Sessions 2.11, 2.12, 2.13, 2.14 and 2.15.
Session 2.11 - Feature engineering
This session explained basic concepts of feature engineering.
Key takeaways:
- Feature engineering involves creating new features, potentially better features, from the currently available features.
Session 2.12 - Categorical variables
This session explained One-hot encoding to use categorical features in the model.
Key takeaways:
- Categorical features cannot be directly used in the model.
- Some features look numerical but can be categorical by nature. Example: number of doors in a car.
Session 2.13 - Regularization
This session gave an introduction to regularization.
Key takeaways:
- To apply regularization α in a normal equation add α to the diagonal elements of matrix XTX.
Session 2.14 - Tuning the model
This session introduced grid search for tuning the model.
Session 2.15 - Using the model
This session showed how to re-train the final model on training and validation set, test it using test set and make predictions using the model.
Estimated Time Taken: 30 minutes