Course
Skills
Feature Engineering
Welcome to Feature Engineering where we will discuss good vs. bad features and how you can preprocess and transform them for optimal use in your models.
What you'll learn
Want to know how you can improve the accuracy of your ML models? What about how to find which data columns make the most useful features? Welcome to Feature Engineering where we will discuss good vs. bad features and how you can preprocess and transform them for optimal use in your models.
Table of contents
Introduction
1min
Raw data to features
55mins
- Raw Data to Features 3m
- Good vs Bad Features 3m
- Quiz: Features are Related to the Objective 3m
- Quiz: Features are knowable at prediction time 4m
- Features are knowable at prediction time' 3m
- Features should be numeric 0m
- Quiz: Features should be numeric 5m
- Features should have enough examples 1m
- Quiz: Features should have enough examples (part 1) 2m
- Quiz: Features should have enough examples (part 2) 3m
- Bringing human insights 0m
- Representing Features 9m
- ML vs Statistics 3m
- Getting Started With GCP And Qwiklabs 4m
- Lab: [ML on GCP C4] Improving model accuracy with new features 0m
- Improve model accuracy with new features 12m
Preprocessing and feature creation
53mins
- Preprocessing and feature creation 7m
- Apache Beam / Cloud Dataflow 10m
- A Simple Dataflow Pipeline 0m
- Lab: [ML on GCP C4] A simple Dataflow pipeline (Python) 0m
- Lab Solution: A Simple Dataflow Pipeline 7m
- Data Pipelines at Scale 6m
- MapReduce in Dataflow 1m
- Lab: [ML on GCP C4] MapReduce in Dataflow (Python) 0m
- Lab Solution: MapReduce in Dataflow 4m
- Dataflow Wrapup 0m
- Preprocessing with Cloud Dataprep 7m
- Lab Intro: Computing Time-Windowed Features in Cloud Dataprep 10m
- Lab: [ML on GCP C4] Computing Time-Windowed Features in Cloud Dataprep 0m
- Lab Solution: Computing Time-Windowed Features in Cloud Dataprep 1m
Feature crosses
91mins
- Introduction 1m
- What is a feature cross? 6m
- Discretization 2m
- Memorization vs. Generalization 5m
- Taxi colors 5m
- Lab Intro: Feature Crosses to create a good classifier 0m
- Lab Solution: Feature Crosses to create a good classifier 6m
- Sparsity + Quiz 6m
- Lab Intro: Too Much of a Good Thing 1m
- Lab Solution: Too Much of a Good Thing 7m
- Implementing Feature Crosses 5m
- Embedding Feature Crosses 9m
- Where to Do Feature Engineering 7m
- Feature Creation in TensorFlow 3m
- Feature Creation in DataFlow 3m
- Lab Intro: Improve ML Model with Feature Engineering 1m
- Lab: [ML on GCP C4] Improve ML model with Feature Engineering 0m
- Debrief: ML Fairness 4m
- Solution: Improve ML Model with Feature Engineering 20m
TensorFlow Transform
43mins
Summary
4mins
