Communications of the ACM
Apache Moves Forward With Machine-Learning Project
Apache Lucene developers have released version 0.2 of the Apache Mahout project, which is an attempt to build libraries and algorithms needed to create machine-learning applications. Version 0.2 offers developers building blocks for creating systems such as the NetFlix recommendation system. The developers say they decided to launch the Apache Mahout project because they believe machine learning needs a strong community, a scalable solution, and a business-friendly model for licensing tools.
The core algorithms involve clustering and categorization, according to Grant Ingersoll, one of the co-founders of the Mahout project. "We also added in frequent pattern mining and a big chunk of code around collaborative filtering, what some people call recommendation systems," Ingersoll says. "The frequent pattern mining is a variation of association rules, such as the proverbial 'when someone buys diapers they also buy beer' scenario."
Apache Hadoop will serve as the most common platform for building with Mahout.
From SD Times
View Full Article
Abstracts Copyright © 2009 Information Inc., Bethesda, Maryland, USA 
No entries found