calculate m automatically

2016-12-18 15:10:16 +00:00
parent 83bd36b139
commit 6b31903c02
1 changed files with 2 additions and 2 deletions
@@ -12,9 +12,9 @@ object KMeans {


  def train(dataset : DataFrame, iterations:Int) : Unit = {
-    val K = 10
-    val m = 4
+    val K = 10  //Number of clusters
    val relevantData = dataset.select("Reputation", "Views", "UpVotes", "DownVotes")
+    val m = relevantData.columns.length  //Number of features
    val rows = relevantData.rdd
    val rowsAsArray = rows.map(row => convertRow(row, m)).persist()