How to…

General

• Extending H2O Jar Manually
• Calling H2O Algorithms
• Running Sparkling Water
• Changing the Default Port
• Hive Support in Sparkling Water

Security

• Enabling SSL
• Enabling LDAP
  • Configuring LDAP in Scala
  • Configuring LDAP in Python (PySparkling)
• Enabling Kerberos Authentication
  • Configuring Kerberos Auth in Scala
  • Configuring Kerberos Auth in Python (PySparkling)
• Running Sparkling Water on Kerberized Hadoop Cluster
  • Internal Backend
  • External Backend
• Using SSL to secure H2O Flow
  • Using existing Java keystore
  • Generate the files automatically

Frames Conversions and Creation

• Spark Frame <–> H2O Frame Conversions
  • Converting an H2OFrame into an RDD[T]
  • Converting an H2OFrame into a DataFrame
  • Converting an RDD[T] into an H2OFrame
  • Converting a DataFrame into an H2OFrame
• H2O Frame as Spark’s Data Source
  • Usage in Python - PySparkling
  • Usage in Scala
  • Specifying Saving Mode
• Creating H2OFrame from an Existing Key
• Import & Export H2O Frames from/to S3
  • Specify the AWS Dependencies
  • Configuring S3A
  • Configuring S3N
  • Sparkling Water Example Code
  • PySparkling Example Code

Modelling

• Train XGBoost Model in Sparkling Water
  • XGBoost Memory Configuration
• Train AutoML Model in Sparkling Water
  • Enabling XGBoost Models when Running Sparkling Water in a Distributed Environment (YARN)
• Train KMeans Model in Sparkling Water
  • H2O KMeans Parameters
• Target Encoding in Sparkling Water
  • Parameters
  • Using Target Encoder

Spark Pipelines

• Using Grid Search GBM in Spark Pipelines
  • Prepare the environment
  • Define the Pipeline Stages
  • Create and Train the Pipeline
  • Run Predictions

Logging

• Change Sparkling Shell Logging Level
• Change Sparkling Shell Logs Location
  • Client
  • Worker Nodes
• Obtain Sparkling Water Logs
  • Logs for Sparkling Water on YARN
  • Logs for Standalone Sparkling Water

As Spark Package

• Use Sparkling Water via Spark Packages