You're reading from Machine Learning with BigQuery ML Create, execute, and improve machine learning models in BigQuery using standard SQL queries

Product type Paperback

Published in Jun 2021

Publisher Packt

ISBN-13 9781800560307

Length 344 pages

Edition 1st Edition

Languages

Python

Tools

BigQuery

Concepts

Machine Learning

Author (1):

Alessandro Marrandino

View More author details

Table of Contents (20) Chapters

Preface

1. Section 1: Introduction and Environment Setup

2. Chapter 1: Introduction to Google Cloud and BigQuery FREE CHAPTER

3. Chapter 2: Setting Up Your GCP and BigQuery Environment

4. Chapter 3: Introducing BigQuery Syntax

5. Section 2: Deep Learning Networks

6. Chapter 4: Predicting Numerical Values with Linear Regression

7. Chapter 5: Predicting Boolean Values Using Binary Logistic Regression

8. Chapter 6: Classifying Trees with Multiclass Logistic Regression

9. Section 3: Advanced Models with BigQuery ML

10. Chapter 7: Clustering Using the K-Means Algorithm

11. Chapter 8: Forecasting Using Time Series

12. Chapter 9: Suggesting the Right Product by Using Matrix Factorization

13. Chapter 10: Predicting Boolean Values Using XGBoost

14. Chapter 11: Implementing Deep Neural Networks

15. Section 4: Further Extending Your ML Capabilities with GCP

16. Chapter 12: Using BigQuery ML with AI Notebooks

17. Chapter 13: Running TensorFlow Models with BigQuery ML

18. Chapter 14: BigQuery ML Tips and Best Practices

19. Other Books You May Enjoy

Summary

In this chapter, we've implemented two DNNs. We remembered the business scenario that was already introduced in Chapter 4, Predicting Numerical Values with Linear Regression. The use case was based on the need to predict the rental time for the New York City bike-sharing service. After that, we learned the basics of DNNs and the different activation functions that can be used to implement the neurons in a network.

We segmented the BigQuery public dataset into three different tables: one to host training data, the second one for the evaluation stage, and the last one to test our DNN model.

During the training phase of the BigQuery ML model, we tested two different activation functions, ReLU and CReLU, comparing the mean absolute error to find the best one.

After that, we evaluated our DNN models on a new set of records to prevent any overfitting and get more confident about the good quality of our BigQuery ML models.

Finally, we applied the model, based on the...