Machine Learning with SAP

Models and Applications

Laboni Bhowmik, Avijit Dhar, Ranajay Mukherjee

• Publisher: SAP PRESS
• Year: 2020
• Print ISBN: 9781493219261
• eBook ISBN: 9781493219278
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Science General
• Index: Computer Science

Work smarter with machine learning! Begin with core machine learning concepts—types of learning, algorithms, data preparation, and more. Then use SAP Data Intelligence, SAP HANA, and other technologies to create your own machine learning applications. Master the SAP HANA Predictive Analysis Library (PAL) and machine learning functional and business services to train and deploy models. Finally, see machine learning in action in industries from manufacturing to banking. a. Foundation
Build your understanding of probability concepts and algorithms that drive machine learning. See how linear regression, classification, and cluster analysis algorithms work, before plugging them into your very own machine learning app!
b. Development
Follow step-by-step instructions to gather and prepare data, create machine learning models, train and fine-tune models, and deploy your final app, all using SAP HANA and SAP Data Intelligence.
c. Platforms
Use built-in SAP HANA libraries to create applications that consume machine learning algorithms or integrate with the R language for additional statistical capabilities. Work with the SAP Leonardo functional services to customize and embed pre-trained models into applications or bring your own model with the help of Google TensorFlow. 1) Development
2) Retraining
3) Implementation
4) SAP Data Intelligence
5) SAP HANA predictive analysis library
6) SAP HANA extended machine learning library
7) SAP HANA automated predictive library
8) Google TensorFlow
9) Embedded machine learning
10) SAP Conversational AI
11) SAP Analytics Cloud Smart Predict

Part I

Introduction

1 Machine Learning and Intelligent Enterprise

In this chapter, we’ll provide a detailed introduction to machine learning and show you how to leverage machine learning to build intelligent and integrated enterprise applications. With these applications, you can automate day-to-day business processes and improve your interactions with customers, suppliers, and collaborators.

An intelligent enterprise is the future that every business dreams of. But before we ride into the future, let’s first take a step back and look at the past, revisiting all footsteps we’ve taken so far:

• 1960s–1980s, mainframe and personal computers
  An automation revolution driven by the innovation of transistors and silicon chips along with emergence of large-scale mainframe system and personal computers.
• 1990s–2000s, client server and internet
  Demonstrated widespread adoption of personal computers, broadband internet, rise of enterprise resource planning (ERP) systems and process-oriented technologies.
• 2000s–2010s, cloud, mobile, and big data
  Digital transformation, it’s all about smartphones, cloud computing, social networks, and big data.
• 2010s–2020s, intelligent technolog ies
  Taking advantage of technological advancements in the areas of machine learning, artificial intelligence, Internet of Things, blockchain, and advance analytics including prediction.

Over the last few decades, technology has evolved significantly: We’ve gone from the era of industrial automation to business process automation to digital transformation and finally arrived at the era of the intelligent enterprise. The era of industrial automation was all about transistors and the silicon chip revolution along with the emergence of large-scale mainframe systems and PCs focusing on shop floor automation. Meanwhile, the era of business process automation demonstrated the widespread adoption of PCs, broadband Internet, enterprise resource planning (ERP) systems, and process-oriented technologies. Then, we entered the era of digital transformation, which is all about smartphones, cloud computing, social networks, and big data. The final but the most anticipated era is the current era of enterprise intelligence, taking advantage of technological advancements in the area of machine learning, artificial intelligence (AI), Internet of Things (IoT), blockchain, and advanced analytics including the prediction of events before they occur.

For many decades, companies have invested heavily on digitally transforming their business processes and related operations, and the digital transformation so far has helped these organizations use their data with increasing visibility and agility and make more insightful decisions. But, surprisingly, many of these organizations have started realizing that digital transformation alone is not enough to survive and must become intelligent enterprise in today’s era. The intelligent enterprise, shown in Figure 1.1, is where business processes and customer experiences are driven by intelligent applications that can make informed decisions and navigate flexible operational processes.

Figure 1.1 The Intelligent Enterprise

In brief, an intelligent enterprise typically facilitates the following advantages:

• People can work seamlessly and access any systems from any device at any time.
• Systems are capable of prioritizing work for people.
• Solutions for dealing with business problems are proposed.
• Repetitive and time-consuming tasks are performed in an automated fashion.
• Systems can discover new patterns and insights from business data during processing, learning from this data to make smarter decisions at the right times.

We’ll introduce you to machine learning in Section 1.1 and walk you through a brief history of how we’ve arrived at the intelligent era in Section 1.2. Then, you’ll learn all about numerous intelligent enterprise use cases in Section 1.3, before we focus our attention on the powerful machine learning portfolio offered by SAP in Section 1.5.

1.1 What Is Machine Learning?

Machine learning is a form of AI that enables a system to learn from data rather than through explicit programming. The precision delivered by a machine learning model is completely dependent on its training data—the more training data you use, the more accurate your predictions. A machine learning model basically represents the output generated when you train your machine learning algorithm with data. After training, when you provide a model with an input, you’ll be given an output. For example, a predictive algorithm will create a predictive model. Then, when you provide the predictive model with data, you’ll receive a prediction based on the data that trained the model.

A key force behind enterprise digital transformation these days, machine learning has been gaining lot of momentum, and enterprises around the world are taking advantage of it instead of using traditional software-based rules to uncover unknown patterns, analyses, and facts from various sources (for instance, text, images, social feed, big data, and IoT devices) including provisions for “training” and “learning” models. Along with SAP, some key key players in this domain include IBM Watson cognitive services, Google TensorFlow, and Azure Machine Learning, which have been operational for a long time.

1.2 Transition from the Digital Era to the Intelligent Era

The digital era of the last decade has evolved into today’s intelligent era by taking advantage of technological advancements especially in the area of artificial intelligence (AI) and machine learning. Using these capabilities, your organization can automate manual activities, identify potential risks, and stay way ahead of the competition. To maximize the gains from technological advancements and minimize disruption, today’s organizations must evolve into an intelligent enterprise that can connect business processes with advanced technologies and software solutions and can also quickly adapt to changing business conditions.

Machine learning, as the basis for intelligent applications in the intelligent enterprise, has emerged as one of the hottest buzzwords as enterprises around the world discover the potential of machine learning to uncover new dimensions of analysis a...

Table of contents

• Dear Reader
• Notes on Usage
• Table of Contents
•   Preface
• Part I   Introduction
• 1   Machine Learning and Intelligent Enterprise
• 2   Machine Learning Fundamentals
• 3   Implementation Lifecycle
• 4   Machine Learning on SAP HANA
• 5   Machine Learning with SAP Data Intelligence
• Part II   Building Machine Learning Applications
• 6   SAP HANA Predictive Analysis Library
• 7   Developing Applications with SAP HANA Predictive Analysis Library
• 8   SAP AI Business Services
• 9   Building Scenarios Using Jupyter Notebook
• 10   Automated Machine Learning Data Science Automation
• 11   Conversational Artificial Intelligence
• Part III   Use Cases and Roadmaps
• 12   Integrating Machine Learning with the Internet of Things
• 13   Industry Use Cases for Machine Learning Applications
• 14   Conclusion and Roadmap
• The Authors
• Index
• Service Pages
• Legal Notes