H2O Instructions

Instructions to get started on the WeCode H2O Workshop.

In this workshop R will only be a requirement if you want to follow the R code, otherwise you can still do the workshop with the H2O Flow feature, you won't need any R dependency for that.

H2O Flow is a feature which lets you do some basic ML directly from your browser with a Java backend, so please be sure to install Java.

Requirements

1. Operating Systems:

• Windows 7 or later
• OS X 10.9 or later
• Ubuntu 12.04
• RHEL/CentOS 6 or later

2. Languages:

• Java is always required. Supported versions include:
  • Java 7 or later. Note: Java 9 is released, but is not currently supported.
• To run the H2O binary using either the command line, R, or Python packages, only 64-bit JRE is required.

3. Browser:

• An internet browser is required to use H2O’s web UI, Flow.
• Latest version of Chrome, Firefox, Safari, or Internet Explorer.

Install Java Dependecies

You can find a download file in the Oracle Website.

If you are using OSX with Homebrew run

$ brew update
$ brew tap caskroom/cask
$ brew install Caskroom/cask/java

Install R

Download R directly from CRAN.

Install RStudio

RStudio makes R easier to use. It includes a code editor, debugging & visualization tools. Download the free version.

Be aware RStudio is only an IDE, you need to install R in order to use RStudio.

Install H2O

To use it standalone (without R)

Install H2O last stable version directly from their website.

R Package

To install the R package run the following code and you'll get a clean version of the latest package:

# The following two commands remove any previously installed H2O packages for R.
if ("package:h2o" %in% search()) { detach("package:h2o", unload=TRUE) }
if ("h2o" %in% rownames(installed.packages())) { remove.packages("h2o") }

# Next, we download packages that H2O depends on.
pkgs <- c("RCurl","jsonlite")
for (pkg in pkgs) {
  if (! (pkg %in% rownames(installed.packages()))) { install.packages(pkg) }
}

# Now we download, install and initialize the H2O package for R.
install.packages("h2o", type="source", repos="http://h2o-release.s3.amazonaws.com/h2o/rel-wolpert/1/R")

# For more: http://h2o-release.s3.amazonaws.com/h2o/rel-lambert/5/docs-website/Ruser/Rinstall.html

Flow

Docs

Download H2O

cd ~/Downloads/h2o-{version}
java -jar h2o.jar

Check your dependencies

• Check your Java version with $ java -version from the command line.

java version "1.8.0_144"

• Check your R version with $ R --version form the command line or opening RStudio and taking a look at the console.

R version 3.4.3 (2017-11-30)

• Check your H2O installation with $ java -jar h2o.jar from the command line.

• Check your R H2O package with from the R console.

library(h2o)
h2o.init()

Data

We will try to cover two ML algorithms. For that we will use two different data sets (last one is a stopwords dataset to improve the algorithm), please be sure to download them before the workshop:

• Airlines data
• Amazon Reviews
• Stopwords for NLP

Code

Repository