Hands-On Machine Learning with Scikit-Learn and TensorFlow

Table of Contents
1.
The Machine Learning Landscape. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
What Is Machine Learning? 10
Why Use Machine Learning? 10
Types of Machine Learning Systems 13
Supervised/Unsupervised Learning 14
Batch and Online Learning 21
Instance-Based Versus Model-Based Learning 24
Main Challenges of Machine Learning 30
Insufficient Quantity of Training Data 30
Nonrepresentative Training Data 32
Poor-Quality Data 33
Irrelevant Features 33
Overfitting the Training Data 34
Underfitting the Training Data 36
Stepping Back 36
Testing and Validating 37
Exercises 39
2.
End-to-End Machine Learning Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Working with Real Data 41
Look at the Big Picture 43
Frame the Problem 43
Select a Performance Measure 45
Check the Assumptions 48
Get the Data 48
Create the Workspace 48
Download the Data 51
Take a Quick Look at the Data Structure 53
iii

---

Create a Test Set 57
Discover and Visualize the Data to Gain Insights 61
Visualizing Geographical Data 62
Looking for Correlations 64
Experimenting with Attribute Combinations 67
Prepare the Data for Machine Learning Algorithms 68
Data Cleaning 69
Handling Text and Categorical Attributes 71
Custom Transformers 73
Feature Scaling 74
Transformation Pipelines 75
Select and Train a Model 77
Training and Evaluating on the Training Set 77
Better Evaluation Using Cross-Validation 78
Fine-Tune Your Model 81
Grid Search 81
Randomized Search 83
Ensemble Methods 84
Analyze the Best Models and Their Errors 84
Evaluate Your System on the Test Set 85
Launch, Monitor, and Maintain Your System 86
Try It Out! 87
Exercises 87