Book Image

Machine Learning with BigQuery ML

By : Alessandro Marrandino
Book Image

Machine Learning with BigQuery ML

By: Alessandro Marrandino

Overview of this book

BigQuery ML enables you to easily build machine learning (ML) models with SQL without much coding. This book will help you to accelerate the development and deployment of ML models with BigQuery ML. The book starts with a quick overview of Google Cloud and BigQuery architecture. You'll then learn how to configure a Google Cloud project, understand the architectural components and capabilities of BigQuery, and find out how to build ML models with BigQuery ML. The book teaches you how to use ML using SQL on BigQuery. You'll analyze the key phases of a ML model's lifecycle and get to grips with the SQL statements used to train, evaluate, test, and use a model. As you advance, you'll build a series of use cases by applying different ML techniques such as linear regression, binary and multiclass logistic regression, k-means, ARIMA time series, deep neural networks, and XGBoost using practical use cases. Moving on, you'll cover matrix factorization and deep neural networks using BigQuery ML's capabilities. Finally, you'll explore the integration of BigQuery ML with other Google Cloud Platform components such as AI Platform Notebooks and TensorFlow along with discovering best practices and tips and tricks for hyperparameter tuning and performance enhancement. By the end of this BigQuery book, you'll be able to build and evaluate your own ML models with BigQuery ML.
Table of Contents (20 chapters)
1
Section 1: Introduction and Environment Setup
5
Section 2: Deep Learning Networks
9
Section 3: Advanced Models with BigQuery ML
15
Section 4: Further Extending Your ML Capabilities with GCP

Drawing business conclusions

In this section, we'll apply our DNN model and understand how many times the BigQuery ML model is able to predict a rental duration close to the actual one.

We'll add a parent SELECT COUNT statement to the previous query to count how many times the difference between the actual duration and the predicted one is less than 15 minutes.

Let's execute the following query to calculate how often the trip duration predictions are far from the actual values:

SELECT COUNT (*)
FROM (
SELECT
   tripduration as actual_duration,
   predicted_label as predicted_duration,
   ABS(tripduration - predicted_label) difference_in_min
FROM
  ML.PREDICT(MODEL  `11_nyc_bike_sharing_dnn.trip_duration_by_stations_day_age_relu`,
    (
    SELECT
          start_station_name,
        ...