[CONF] Apache Mahout > Mailing Lists, IRC and Archives

[shantha kumar N (Confluence)]

Mailing Lists, IRC and Archives Page comment added by shantha kumar N Hi, in Mahout examples, the (org.apache.mahout.classifier.sgd.)RunLogistic or Trainlogistic class is a great example to classify content with SGD and to get a nice confusion matrix. I'm trying to use and adapt this to classify data in more than 2 categories. The alorithm uses classify scalar methods, which is only working with 2 categories. "classifyScalar" method in AbstractOnlineLogisticRegression class return a double when we evaluate the model. This double is used as a score fot the matrix. "classify" seems to be equivalent but returns another vector, I guess because the matrix has more than one row in that case. I'm looking for a way to do this with more categories. Do you have a clue or example for doing this i mean --categories 4 is possible instead of --categories 2 in the below command.If not then how to SGD algorithm for multinomial logistic regression in mahout. bin/mahout trainlogistic --input donut.csv \ --output ./model \ --target color --categories 4 \ --predictors x y --types numeric \ --features 20 --passes 100 --rate 50 Stop watching space | Change email notification preferences View Online | Reply To This