streaming-matrix-factorization (homepage)

Streaming Recommendation Engine using matrix factorization with user and product bias

@brkyvz / (2)

Matrix Factorization Model for Recommendation Systems. The model consists of
- user factors (User Matrix, `U`),
- product factors (Product Matrix, `P^T^`),
- user biases (user bias vector, `bu`),
- product biases (product bias vector, `bp`) and
- the global bias (global average, `mu`).
A rating is predicted as:
r = U(u) * P^T^(i) + bu(u) + bp(i) + mu
Can be trained both on a stream or a single RDD and predictions can be made to a stream
or a single RDD.


Tags

  • 1|ml
  • 1|streaming
  • 1|machine learning
  • 1|mllib

How to

Include this package in your Spark Applications using:

spark-shell, pyspark, or spark-submit

> $SPARK_HOME/bin/spark-shell --packages brkyvz:streaming-matrix-factorization:0.1.0

sbt

If you use the sbt-spark-package plugin, in your sbt build file, add:

spDependencies += "brkyvz/streaming-matrix-factorization:0.1.0"

Otherwise,

resolvers += "Spark Packages Repo" at "http://dl.bintray.com/spark-packages/maven"

libraryDependencies += "brkyvz" % "streaming-matrix-factorization" % "0.1.0"

Maven

In your pom.xml, add:
<dependencies>
  <!-- list of dependencies -->
  <dependency>
    <groupId>brkyvz</groupId>
    <artifactId>streaming-matrix-factorization</artifactId>
    <version>0.1.0</version>
  </dependency>
</dependencies>
<repositories>
  <!-- list of other repositories -->
  <repository>
    <id>SparkPackagesRepo</id>
    <url>http://dl.bintray.com/spark-packages/maven</url>
  </repository>
</repositories>

Releases

Version: 0.1.0 ( edfdbf | zip | jar ) / Date: 2015-05-26 / License: Apache-2.0 / Scala version: 2.10

Spark Scala/Java API compatibility: - 100% , - 100% , - 100% , - 100%