book

Machine Learning at Scale with H2O

Name: Machine Learning at Scale with H2O
ISBN: 9781800566019

by Gregory Keys, David Whiting

July 2022

Beginner to intermediate

396 pages

8h 17m

English

Packt Publishing

Read now

Unlock full access

Machine Learning at Scale with H2O
AcknowledgmentsContributorsAbout the authorsAbout the reviewers
Preface
Who this book is forWhat this book coversTo get the most out of this bookDownload the example code filesDownload the color imagesConventions usedGet in touchShare Your Thoughts
Section 1 – Introduction to the H2O Machine Learning Platform for Data at Scale
Chapter 1: Opportunities and Challenges
ML at scaleThe ML life cycle and three challenge areas for ML at scaleA simplified ML life cycleThe model building challenge – state-of-the-art models at scaleThe business challenge – getting your models into enterprise production systemsThe navigation challenge – navigating the enterprise stakeholder landscapeH2O.ai's answer to these challengesSummary
Chapter 2: Platform Components and Key Concepts
Technical requirementsHello World – the H2O machine learning codeCode exampleSome issues of scaleThe components of H2O machine learning at scaleH2O Core – in-memory distributed model buildingH2O Enterprise Steam – a managed, self-provisioning portalThe H2O MOJO – a flexible, low-latency scoring artifactThe workflow using H2O componentsH2O key conceptsThe data scientist's experienceThe H2O clusterEnterprise Steam as an H2O gatewayEnterprise Steam and the H2O Core high-level architectureSparkling Water allows users to code in H2O and Spark seamlessly MOJOs export as DevOps-friendly artifactsSummary
Chapter 3: Fundamental Workflow – Data to Deployable Model
Technical requirementsUse case and data overviewThe fundamental workflowStep 1 – launching the H2O clusterStep 2 – connecting to the H2O cluster Step 3 – building the modelStep 4 – evaluating and explaining the modelStep 5 – exporting the model's scoring artifactStep 6 – shutting down the clusterVariation points – alternatives and extensions to the fundamental workflowLaunching an H2O cluster using the Enterprise Steam API versus the UI (step 1)Launching an H2O-3 versus Sparkling Water cluster (step 1)Implementing Enterprise Steam or not (steps 1–2)Using a personal access token to log in to Enterprise Steam (step 2)Building the model (step 3)Evaluating and explaining the model (step 4)Exporting the model's scoring artifact (step 5)Shutting down the cluster (step 6)Summary
Section 2 – Building State-of-the-Art Models on Large Data Volumes Using H2O
Chapter 4: H2O Model Building at Scale – Capability Articulation
H2O data capabilities during model buildingIngesting data from the source to the H2O clusterManipulating data in the H2O clusterExporting data out of the H2O clusterAdditional data capabilities provided by Sparkling WaterH2O machine learning algorithmsH2O unsupervised learning algorithmsH2O supervised learning algorithmsParameters and hyperparametersH2O extensions of supervised learningMiscellaneousH2O modeling capabilitiesH2O model training capabilitiesH2O model evaluation capabilitiesH2O model explainability capabilitiesH2O trained model artifactsSummary
Chapter 5: Advanced Model Building – Part I
Technical requirementsSplitting data for validation or cross-validation and testingTrain, validate, and test set splitsTrain and test splits for k-fold cross-validationAlgorithm considerationsAn introduction to decision treesRandom forestsGradient boostingBaseline model trainingModel optimization with grid searchStep 1 – a Cartesian grid search to focus on the best tree depthStep 2 – a random grid search to tune other parametersH2O AutoMLThe AutoML leaderboardFeature engineering optionsTarget encodingOther feature engineering optionsLeveraging H2O Flow to enhance your IDE workflowMonitoring with FlowInteractive investigations with FlowPutting it all together – algorithms, feature engineering, grid search, and AutoMLAn enhanced AutoML procedureSummary
Chapter 6: Advanced Model Building – Part II
Technical requirementsModeling in Sparkling WaterIntroducing Sparkling Water pipelinesImplementing a sentiment analysis pipelineImporting the raw Amazon dataDefining Spark pipeline stagesCreating a Sparkling Water pipelineLooking ahead – a production previewUL methods in H2OWhat is anomaly detection?Isolation forests in H2OBest practices for updating H2O modelsRetraining modelsCheckpointing modelsEnsuring H2O model reproducibilityCase 1 – Reproducibility in single-node clustersCase 2 – Reproducibility in multi-node clustersReproducibility for specific algorithms Best practices for reproducibilitySummary

Chapter 7: Understanding ML Models
Selecting model performance metricsExplaining models built in H2OA simple introduction to Shapley valuesGlobal explanations for single modelsLocal explanations for single modelsGlobal explanations for multiple modelsAutomated model documentation (H2O AutoDoc)Summary
Chapter 8: Putting It All Together
Technical requirementsData wranglingImporting the raw dataDefining the problem and creating the response variableConverting implied numeric data from strings into numeric valuesCleaning up messy categorical columnsFeature engineeringAlgebraic transformationsFeatures engineered from datesSimplifying categorical variables by combining categories Missing value indicator functionsTarget encoding categorical columnsModel building and evaluationModel search and optimization with AutoML Investigating global explainability with AutoML modelsSelecting a model from the AutoML candidatesFinal model evaluationPreparation for model pipeline deployment Summary
Section 3 – Deploying Your Models to Production Environments
Chapter 9: Production Scoring and the H2O MOJO
Technical requirementsThe model building and model scoring contextsModel training to production model scoringH2O production scoringEnd-to-end production scoring pipeline with H2OTarget production systems for H2O MOJOsH2O MOJO deep diveWhat is a MOJO?Deploying a MOJOWrapping MOJOs using the H2O MOJO APIObtaining the MOJO runtimeThe h2o-genmodel APIA generalized approach to wrapping your MOJO Wrapping example – Build a batch file scorer in JavaOther things to know about MOJOsInspecting MOJO decision logicMOJO and POJOSummary
Chapter 10: H2O Model Deployment Patterns
Technical requirementsSurveying a sample of MOJO deployment patternsH2O softwareThird-party software integrationsYour software integrationsAccelerators based on H2O Driverless AI integrationsExploring examples of MOJO scoring with H2O software H2O MLOpsH2O eScorer H2O batch database scorerH2O batch file scorerH2O Kafka scorerH2O batch scoring on SparkExploring examples of MOJO scoring with third-party softwareSnowflake integrationTeradata integrationBI tool integrationUiPath integrationExploring examples of MOJO scoring with your target-system softwareYour software applicationOn-device scoringExploring examples of accelerators based on H2O Driverless AI integrations Apache NiFiApache FlinkAWS LambdaAWS SageMakerSummary
Section 4 – Enterprise Stakeholder Perspectives
Chapter 11: The Administrator and Operations Views
A model building and deployment view – the personas on the groundView 1 – Enterprise Steam administratorEnterprise Steam administrator concerns Enterprise Steam configurationsH2O user governance from Enterprise SteamEnterprise Steam configurationsServer cluster (backend) integrationH2O-3 and Sparkling Water managementRestarting Enterprise SteamView 2 – The operations teamEnterprise Steam server OpsH2O cluster OpsMLOpsView 3 – The data scientistInteractions with Enterprise Steam administratorsInteractions with H2O cluster (Hadoop or Kubernetes) Ops teamsInteractions with MLOps teamsSummary
Chapter 12: The Enterprise Architect and Security Views
Technical requirementsThe enterprise and security architect viewH2O at Scale enterprise architectureH2O at Scale implementation patternsComponent integration architectureCommunication architectureDeployment architectureH2O at Scale securityData movement and privacyUser authentication and access controlNetwork and firewallThe data scientist's view of enterprise and security architectureSummary
Section 5 – Broadening the View – Data to AI Applications with the H2O AI Cloud Platform
Chapter 13: Introducing H2O AI Cloud
Technical requirementsAn H2O AI Cloud overviewH2O AI Cloud component breakdownDistributedML (H2O-3 and H2O Sparkling Water)H2O AutoML (H2O Driverless AI)DeepLearningML (H2O Hydrogen Torch)DocumentML (H2O Document AI)A self-provisioning service (H2O Enterprise Steam)Feature Store (H2O AI Feature Store)MLOps (H2O MLOps)Low-code SDK for AI applications (H2O Wave)App Store (H2O AI App Store)H2O AI Cloud architectureSummary
Chapter 14: H2O at Scale in a Larger Platform Context
Technical requirementsA quick recap of H2O AI CloudExploring a baseline reference solution for H2O at scale Exploring new possibilities for H2O at scale Leveraging H2O Driverless AI for prototyping and feature discoveryIntegrating H2O MLOps for model monitoring, management, and governanceLeveraging H2O AI Feature Store for feature operationalization and reuseConsuming predictions in a business context from a Wave AI appIntegrating an automated retraining pipeline in a Wave AI appA Reference H2O Wave app as an enterprise AI integration fabricSummary
Appendix : Alternative Methods to Launch H2O Clusters
Local H2O-3 clusterStep 1 – Install H2O-3 in PythonStep 2 – Launch your H2O-3 cluster and write codeLocal Sparkling Water clusterStep 1 – Install Spark locallyStep 2 – Install Sparkling Water in Python Step 3 – Install a Sparkling Water Python interactive shellStep 4 – Launch a Jupyter notebook on top of the Sparkling Water shellStep 5 – Launch your Sparkling Water cluster and write codeH2O-3 cluster in the 90-day free trial environment for H2O AI CloudStep 1 – Get your 90-day trial to H2O AI CloudStep 2 – Set up your Python environmentStep 3 – Launch your clusterStep 4 – Write H2O-3 code
Why subscribe?
Other Books You May EnjoyPackt is searching for authors like youShare Your Thoughts

Content preview from Machine Learning at Scale with H2O

Chapter 10: H2O Model Deployment Patterns

In the previous chapter, we learned how easy it is to generate a ready-to-deploy scoring artifact from our model-building step and how this artifact, called a MOJO, is designed to flexibly deploy to a wide diversity of production systems.

In this chapter, we explore this flexibility of MOJO deployment by surveying a wide range of MOJO deployment patterns and digging down into the details of each deployment pattern. We will see how MOJOs are implemented for scoring on either H2O software, third-party software including business intelligence (BI) tools, and your own software. These implementations will include scoring on real-time, batch, and streaming data.

Recall from Chapter 1, Opportunities and Challenges ...

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 9781800566019

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

Machine Learning at Scale with H2O

by Gregory Keys, David Whiting

Chapter 10: H2O Model Deployment Patterns

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.