Book Image

Automated Machine Learning with Microsoft Azure

By : Dennis Michael Sawyers
Book Image

Automated Machine Learning with Microsoft Azure

By: Dennis Michael Sawyers

Overview of this book

Automated Machine Learning with Microsoft Azure will teach you how to build high-performing, accurate machine learning models in record time. It will equip you with the knowledge and skills to easily harness the power of artificial intelligence and increase the productivity and profitability of your business. Guided user interfaces (GUIs) enable both novices and seasoned data scientists to easily train and deploy machine learning solutions to production. Using a careful, step-by-step approach, this book will teach you how to use Azure AutoML with a GUI as well as the AzureML Python software development kit (SDK). First, you'll learn how to prepare data, train models, and register them to your Azure Machine Learning workspace. You'll then discover how to take those models and use them to create both automated batch solutions using machine learning pipelines and real-time scoring solutions using Azure Kubernetes Service (AKS). Finally, you will be able to use AutoML on your own data to not only train regression, classification, and forecasting models but also use them to solve a wide variety of business problems. By the end of this Azure book, you'll be able to show your business partners exactly how your ML models are making predictions through automatically generated charts and graphs, earning their trust and respect.
Table of Contents (17 chapters)
Section 1: AutoML Explained – Why, What, and How
Section 2: AutoML for Regression, Classification, and Forecasting – A Step-by-Step Guide
Section 3: AutoML in Production – Automating Real-Time and Batch Scoring Solutions

Installing the many models solution accelerator

The MMSA was built by Microsoft in 2019 to address the needs of a growing number of customers who wanted to train hundreds of thousands of similar ML models simultaneously. This is particularly important for product demand forecasting, where you are trying to make forecasts for many different products at many different locations.

The impetus for the accelerator is model accuracy. While you could train a single model to predict product demand across all of your product lines and all of your stores, you will find that training individual models for each combination of product and store tends to yield superior performance. This is because a multitude of factors are dependent on both your algorithm and your data. It can be very difficult for some algorithms to find meaningful patterns when you're dealing with hundreds of thousands of different products distributed across the globe.

Additionally, the same columns can have different...