Dear All
I tried the Mahout Naive Bayes Classifier with 1300 'Good' document and 1300
'Bad' documents
The code which I used for training is given below.
public void trainerMahout(String inpDir, String outputDir) throws
Exception {
BayesParameters params = new BayesParameters(2);// N-Gram size set to 3
params.set("alpha_i", "1.0");
// params.set("dataSource", "hbase");
params.set("dataSource", "hdfs");
// setting the BayesParameters *compulsory*
TrainClassifier.trainNaiveBayes(inpDir, outputDir, params);
System.out.println("Finished Training");
// Training the system
}
When I tried the model against the development 1300 'Good' it is ok. But the
'Bad' 1300 document also the model classfied as 'Good document'. What may be
the reason
--
**********************************
JAGANADH G
http://jaganadhg.freeflux.net/blog
