The full operation of the mahout is still supported by Hadoop, but many of the algorithms only need to be able to add the Hadoop jar package to classpath to work properly.
For example, when we use logisticmodelparameters , we refer to the package
Java code
- Import org.apache.hadoop.io.Writable;
Import org.apache.hadoop.io.Writable;
According to the previous article, just add the Mahout package to the classpath, you will be prompted to find hadoop***
To solve this problem, it is necessary to build a development environment through MAVEN for project introduction.
Here's how I built the mahout development environment with eclipse in Windows without Hadoop.
How to download and run Eclipse is not in the scope of this article
1. Run Eclipse and select "exsisting maven Projects" from File--Import
2. Select the root directory of the MAHOUT-SRC:
Next, you can next/ok all the way.
You'll need to restart eclipse~ once in the middle.
Finally, in eclipse you will see a list of the following items:
PS: The above steps are valid in Windows, although the import was successful on Centos6.2, but the Java file could not be run properly. At present I also wood has the method ...
Try creating one of the simplest files in Mahout-example:
First run please be patient, the specific reason is unknown (I do not play maven), Eclipse will try to create the Java file Index, look at the performance of the machine, my time is about 5 minutes ~ after all.
Run complete, display:
Hello Classifier!!!
After reading a lot of articles, coupled with their own attempts to feel that this is the easiest way to further use of Mahout: D
Windows + Eclipse Build mahout Run environment