NoSuchMethodError: breeze.math.MutableVectorField.normImplDouble()Lbreeze/generic/UFunc$UImpl2

[Ken Kawamoto]

Hi,

I'm trying to train LogisticClassifier with Example of label Int and features DenseVector[Double], and seeing this error.

[info] Exception encountered when attempting to run a suite with class name: foo.bar.MLTest *** ABORTED ***

[info]   java.lang.NoSuchMethodError: breeze.math.MutableVectorField.normImplDouble()Lbreeze/generic/UFunc$UImpl2;

[info]   at nak.classify.LFMatrix$.space(LFMatrix.scala:313)

[info]   at nak.classify.LogisticClassifier$Trainer.train(LogisticClassifier.scala:72)

I feel I need to import a class to provide the implicit binding, but do now know which one to import.

can you advise me?

below is the code snippet. (I know it's ugly. I'm just testing...)

thanks in advance.

Ken

    val examples: MutableList[Example[Int, DenseVector[Double]]] = new MutableList()

    for (m <- 0 until X.getM) {

      val features: Array[Double] = Array.empty

      for (n <- 0 until X.getN) {

        features :+ X.get(m, n).doubleValue

      }

      examples += Example(y.get(m, 0).toInt, DenseVector(features))

    }

    val opt = new OptParams(

      maxIterations=60,

      useStochastic=false,

      useL1=true)

    val classifier = new LogisticClassifier.Trainer[Int, DenseVector[Double]](opt).train(examples)

[David Hall]

Sorry for the long delay.

You're probably mixing breeze versions, most likely because of the Nak dependency.

[Ken Kawamoto]

thx David. removing breeze libraries from build.sbt (and keeping nak dependency) solves the problem.

—
Sent from Mailbox