IrisPca.scala

/*
IrisPca.scala

 */

import breeze.linalg.*
import breeze.stats.covmat
import scalaglm.*

@main def irisPca() =

    val url = "http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data"
    val fileName = "iris.csv"
    val imap = Map(
      "Iris-setosa" -> 0,
      "Iris-versicolor" -> 1,
      "Iris-virginica" -> 2)

    // download the file to disk if it hasn't been already
    val file = new java.io.File(fileName)
    if !file.exists then
      val s = new java.io.PrintWriter(file)
      val data = scala.io.Source.fromURL(url).getLines()
      data.foreach(l => s.write(l.trim.split(',').
        map(x => imap.getOrElse(x, x)).mkString("", ",", "\n")))
      s.close

    // read the file from disk
    val mat = csvread(file)
    println("Mat Dim: " + mat.rows + " " + mat.cols)
    val x = mat(::, 0 to 3)
    println("X Dim: " + x.rows + " " + x.cols)
    val clas = mat(::, 4).toDenseVector

    println("PCA with built-in Breeze version (like R princomp):")
    val pca = new PCA(x, covmat(x))
    println("Loadings:")
    println(pca.loadings)
    println("Stdev:")
    println(pca.sdev)
    println(pca.scores(0 to 5, ::))

    println("Now my version (like R prcomp):")
    val myPca = Pca(x, List("S-L","S-W","P-L","P-W"))
    println(myPca.loadings) // loadings transposed
    println(myPca.sdev)
    myPca.summary
    println("Scores:")
    println(myPca.scores(0 to 5, ::))
    myPca.plots.saveas("IrisPcaDiag.png")

    // scatter plot first 2 principal components
    import breeze.plot.*
    val fig = Figure("PCA")
    val p = fig.subplot(0)
    val ind0 = (0 until x.rows) filter (i => clas(i) == 0)
    p += plot(myPca.scores(ind0, 0).toDenseVector,
      myPca.scores(ind0, 1).toDenseVector, '.', colorcode = "blue")
    val ind1 = (0 until x.rows) filter (i => clas(i) == 1)
    p += plot(myPca.scores(ind1, 0).toDenseVector,
      myPca.scores(ind1, 1).toDenseVector, '.', colorcode = "red")
    val ind2 = (0 until x.rows) filter (i => clas(i) == 2)
    p += plot(myPca.scores(ind2, 0).toDenseVector,
      myPca.scores(ind2, 1).toDenseVector, '.', colorcode = "green")
    fig.saveas("IrisPca.png")

    // Test without variable names
    Pca(x).summary


// eof