Monday, September 6, 2010

Installing Apache Mahout

To start, you need to install the following prerequisites:

* JDK 1.6 or higher
* Ant 1.7 or higher
* If you want to build the Mahout source, Maven 2.0.9 or 2.0.10

You also need this article's sample code (Download), which includes a copy of Mahout and its dependencies. Follow these steps to install the sample code:

1. unzip sample.zip
2. cd apache-mahout-examples
3. ant install

Step 3 downloads the necessary Wikipedia files and compiles the code. The Wikipedia file used is approximately 2.5 gigabytes, so download times will depend on your bandwidth. It also uncompresses the files. You need to have 10 GB of free space.
Occasionally, there may be an error.
You can try commenting the lines of download of 2.5 GB of wiki pages and download them separately with the link given. Then ant install will give you build failed. Do not worry. A folder named Wikipedia would have been created in the apache-mahout-examples directory. If not, create one. Place the downloaded content of wikipedia file there. Do ant install again. Now those files already downloaded will just uncompress and it will take about half an hour. Depending upon your system configuration, this may vary. The result will be installation of mahout.

No comments:

Post a Comment