Machine Learning for Business: Using Amazon SageMaker and Jupyter`

Machine Learning for Business
brief contents
contents
preface
acknowledgments
about this book
	Who should read this book
	How this book is organized: A roadmap
	About the code
	liveBook discussion forum
about the authors
about the cover illustration
Part 1?Machine learning for business
	1 How machine learning applies to your business
		1.1 Why are our business systems so terrible?
		1.2 Why is automation important now?
			1.2.1 What is productivity?
			1.2.2 How will machine learning improve productivity?
		1.3 How do machines make decisions?
			1.3.1 People: Rules-based or not?
			1.3.2 Can you trust a pattern-based answer?
			1.3.3 How can machine learning improve your business systems?
		1.4 Can a machine help Karen make decisions?
			1.4.1 Target variables
			1.4.2 Features
		1.5 How does a machine learn?
		1.6 Getting approval in your company to use machine learning to make decisions
		1.7 The tools
			1.7.1 What are AWS and SageMaker, and how can they help you?
			1.7.2 What is a Jupyter notebook?
		1.8 Setting up SageMaker in preparation for tackling the scenarios in chapters 2 through 7
		1.9 The time to act is now
		Summary
Part 2?Six scenarios: Machine learning for business
	2 Should you send a purchase order to a technical approver?
		2.1 The decision
		2.2 The data
		2.3 Putting on your training wheels
		2.4 Running the Jupyter notebook and making predictions
			2.4.1 Part 1: Loading and examining the data
			2.4.2 Part 2: Getting the data into the right shape
			2.4.3 Part 3: Creating training, validation, and test datasets
			2.4.4 Part 4: Training the model
			2.4.5 Part 5: Hosting the model
			2.4.6 Part 6: Testing the model
		2.5 Deleting the endpoint and shutting down your notebook instance
			2.5.1 Deleting the endpoint
			2.5.2 Shutting down the notebook instance
		Summary
	3 Should you call a customer because they are at risk of churning?
		3.1 What are you making decisions about?
		3.2 The process flow
		3.3 Preparing the dataset
			3.3.1 Transformation 1: Normalizing the data
			3.3.2 Transformation 2: Calculating the change from week to week
		3.4 XGBoost primer
			3.4.1 How XGBoost works
			3.4.2 How the machine learning model determines whether the function is getting better or getting worse AUC
		3.5 Getting ready to build the model
			3.5.1 Uploading a dataset to S3
			3.5.2 Setting up a notebook on SageMaker
		3.6 Building the model
			3.6.1 Part 1: Loading and examining the data
			3.6.2 Part 2: Getting the data into the right shape
			3.6.3 Part 3: Creating training, validation, and test datasets
			3.6.4 Part 4: Training the model
			3.6.5 Part 5: Hosting the model
			3.6.6 Part 6: Testing the model
		3.7 Deleting the endpoint and shutting down your notebook instance
			3.7.1 Deleting the endpoint
			3.7.2 Shutting down the notebook instance
		3.8 Checking to make sure the endpoint is deleted
		Summary
	4 Should an incident be escalated to your support team?
		4.1 What are you making decisions about?
		4.2 The process flow
		4.3 Preparing the dataset
		4.4 NLP (natural language processing)
			4.4.1 Creating word vectors
			4.4.2 Deciding how many words to include in each group
		4.5 What is BlazingText and how does it work?
		4.6 Getting ready to build the model
			4.6.1 Uploading a dataset to S3
			4.6.2 Setting up a notebook on SageMaker
		4.7 Building the model
			4.7.1 Part 1: Loading and examining the data
			4.7.2 Part 2: Getting the data into the right shape
			4.7.3 Part 3: Creating training and validation datasets
			4.7.4 Part 4: Training the model
			4.7.5 Part 5: Hosting the model
			4.7.6 Part 6: Testing the model
		4.8 Deleting the endpoint and shutting down your notebook instance
			4.8.1 Deleting the endpoint
			4.8.2 Shutting down the notebook instance
		4.9 Checking to make sure the endpoint is deleted
		Summary
	5 Should you question an invoice sent by a supplier?
		5.1 What are you making decisions about?
		5.2 The process flow
		5.3 Preparing the dataset
		5.4 What are anomalies
		5.5 Supervised vs. unsupervised machine learning
		5.6 What is Random Cut Forest and how does it work?
			5.6.1 Sample 1
			5.6.2 Sample 2
		5.7 Getting ready to build the model
			5.7.1 Uploading a dataset to S3
			5.7.2 Setting up a notebook on SageMaker
		5.8 Building the model
			5.8.1 Part 1: Loading and examining the data
			5.8.2 Part 2: Getting the data into the right shape
			5.8.3 Part 3: Creating training and validation datasets
			5.8.4 Part 4: Training the model
			5.8.5 Part 5: Hosting the model
			5.8.6 Part 6: Testing the model
		5.9 Deleting the endpoint and shutting down your notebook instance
			5.9.1 Deleting the endpoint
			5.9.2 Shutting down the notebook instance
		5.10 Checking to make sure the endpoint is deleted
		Summary
	6 Forecasting your company?s monthly power usage
		6.1 What are you making decisions about?
			6.1.1 Introduction to time-series data
			6.1.2 Kiara?s time-series data: Daily power consumption
		6.2 Loading the Jupyter notebook for working with time-series data
		6.3 Preparing the dataset: Charting time-series data
			6.3.1 Displaying columns of data with a loop
			6.3.2 Creating multiple charts
		6.4 What is a neural network?
		6.5 Getting ready to build the model
			6.5.1 Uploading a dataset to S3
			6.5.2 Setting up a notebook on SageMaker
		6.6 Building the model
			6.6.1 Part 1: Loading and examining the data
			6.6.2 Part 2: Getting the data into the right shape
			6.6.3 Part 3: Creating training and testing datasets
			6.6.4 Part 4: Training the model
			6.6.5 Part 5: Hosting the model
			6.6.6 Part 6: Making predictions and plotting results
		6.7 Deleting the endpoint and shutting down your notebook instance
			6.7.1 Deleting the endpoint
			6.7.2 Shutting down the notebook instance
		6.8 Checking to make sure the endpoint is deleted
		Summary
	7 Improving your company?s monthly power usage forecast
		7.1 DeepAR?s ability to pick up periodic events
		7.2 DeepAR?s greatest strength: Incorporating related time series
		7.3 Incorporating additional datasets into Kiara?s power consumption model
		7.4 Getting ready to build the model
			7.4.1 Downloading the notebook we prepared
			7.4.2 Setting up the folder on SageMaker
			7.4.3 Uploading the notebook to SageMaker
			7.4.4 Downloading the datasets from the S3 bucket
			7.4.5 Setting up a folder on S3 to hold your data
			7.4.6 Uploading the datasets to your AWS bucket
		7.5 Building the model
			7.5.1 Part 1: Setting up the notebook
			7.5.2 Part 2: Importing the datasets
			7.5.3 Part 3: Getting the data into the right shape
			7.5.4 Part 4: Creating training and test datasets
			7.5.5 Part 5: Configuring the model and setting up the server to build the model
			7.5.6 Part 6: Making predictions and plotting results
		7.6 Deleting the endpoint and shutting down your notebook instance
			7.6.1 Deleting the endpoint
			7.6.2 Shutting down the notebook instance
		7.7 Checking to make sure the endpoint is deleted
		Summary
Part 3?Moving machine learning into production
	8 Serving predictions over the web
		8.1 Why is serving decisions and predictions over the web so difficult?
		8.2 Overview of steps for this chapter
		8.3 The SageMaker endpoint
		8.4 Setting up the SageMaker endpoint
			8.4.1 Uploading the notebook
			8.4.2 Uploading the data
			8.4.3 Running the notebook and creating the endpoint
		8.5 Setting up the serverless API endpoint
			8.5.1 Setting up your AWS credentials on your AWS account
			8.5.2 Setting up your AWS credentials on your local computer
			8.5.3 Configuring your credentials
		8.6 Creating the web endpoint
			8.6.1 Installing Chalice
			8.6.2 Creating a Hello World API
			8.6.3 Adding the code that serves the SageMaker endpoint
			8.6.4 Configuring permissions
			8.6.5 Updating requirements.txt
			8.6.6 Deploying Chalice
		8.7 Serving decisions
		Summary
	9 Case studies
		9.1 Case study 1: WorkPac
			9.1.1 Designing the project
			9.1.2 Stage 1: Preparing and testing the model
			9.1.3 Stage 2: Implementing proof of concept (POC)
			9.1.4 Stage 3: Embedding the process into the company?s operations
			9.1.5 Next steps
			9.1.6 Lessons learned
		9.2 Case study 2: Faethm
			9.2.1 AI at the core
			9.2.2 Using machine learning to improve processes at Faethm
			9.2.3 Stage 1: Getting the data
			9.2.4 Stage 2: Identifying the features
			9.2.5 Stage 3: Validating the results
			9.2.6 Stage 4: Implementing in production
		9.3 Conclusion
			9.3.1 Perspective 1: Building trust
			9.3.2 Perspective 2: Geting the data right
			9.3.3 Perspective 3: Designing your operating model to make the most of your machine learning capability
			9.3.4 Perspective 4: What does your company look like once you are using machine learning everywhere?
		Summary
Appendix A?Signing up for Amazon AWS
	A.1 Signing up for AWS
	A.2 AWS Billing overview
Appendix B?Setting up and using S3 to store files
	B.1 Creating and setting up a bucket in S3
		B.1.1 Step 1: Naming your bucket
		B.1.2 Step 2: Setting properties for your bucket
		B.1.3 Step 3: Setting permissions
		B.1.4 Step 4: Reviewing settings
	B.2 Setting up folders in S3
	B.3 Uploading files to S3
Appendix C?Setting up and using AWS SageMaker to build a machine learning system
	C.1 Setting up
	C.2 Starting at the Dashboard
	C.3 Creating a notebook instance
	C.4 Starting the notebook instance
	C.5 Uploading the notebook to the notebook instance
	C.6 Running the notebook
Appendix D?Shutting it all down
	D.1 Deleting the endpoint
	D.2 Shutting down the notebook instance
Appendix E?Installing Python
index
	Symbols
	A
	B
	C
	D
	E
	F
	G
	H
	I
	J
	K
	L
	M
	N
	O
	P
	Q
	R
	S
	T
	U
	V
	W
	X
	Z