logoIntroduction toMachine Learnin

Module 5: Preprocessing Numerical Features, Pipelines and Hyperparameter Optimization

In this model, we will concentrate on the steps that need to be taken before building a model. Preparation through imputation and scaling is an important step of model building and can be done using tools such as pipelines. Next, we will explore automated hyperparameter optimization.

0Module Learning Outcomes

1The Importance of Preprocessing

2Questions on Why

3Motivation True and False

4Preprocessing Questions

5Case Study: Preprocessing with Imputation

6Imputation

7Imputation True or False

8Imputing in Action

9Case Study: Preprocessing with Scaling

10Name that Scaling Method!

11Scaling True or False

12Practicing Scaling

13Case Study: Pipelines

14Pipeline Questions

15Pipeline True or False

16Applying Pipelines

17Automated Hyperparameter Optimization

18Exhaustive or Randomized Grid Search

19Hyperparameter Quick Questions

20Using Automated Hyperparameter Optimization in Action!

21What Did We Just Learn?

About this course

This course covers the data science perspective on the introductory concepts in machine learning, with a focus on making predictions. It covers how to build different models such as K-NN, decision trees and linear classifiers as well as important concepts such as data splitting and fundamental rules and laws. In addition, this course will teach you how to evaluate models properly and question their validity all while streamlining the process with pipelines.

About the program

The University of British Columbia (UBC) is a comprehensive research-intensive university, consistently ranked among the 40 best universities in the world. The Key Capabilities in Data Science program was launched in September 2020 and is developed and taught by many of the same instructors as the UBC Master of Data Science program.