I would like to use the Mahout decision tree learning process inference model as the input model for the Weka classifier.
Since training a complex decision tree based on millions of training records is practically impractical for a single Weka node classifier, I would like to use Mahout to build a model using, for example, Random Forest Partial Implementation .
Although the algorithm above may be problematic during training, it is quite simple to use it to predict with Weka on the same machine.
The Mahout wiki site states that data formats for import include the Weka ARFF format, but not for export.
Is it possible to use some of the existing implementations in Mahout to train models that will be used in production with a simple Weka system ?
, , :.arff - , /. , Weka /, Weka java Classifier, Java Serializable. , Java, , JVM. , , , Mahout, Weka / - .
Classifier
Serializable
, ( ~ 45 / ) Weka Random Forest ( 2,26 ), . .