[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » spark » spark-introduction

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision 		Previous revision
Next revision Both sides next revision
spark:spark-introduction [2014/11/03 19:00]
straka 		spark:spark-introduction [2014/11/03 20:31]
straka
Line 99: Line 99:
 </file> </file>
 The implementation starts by loading the data and caching them in memory using ''cache''. Then, standard iterative algorithm is performed, running in parallel.  The implementation starts by loading the data and caching them in memory using ''cache''. Then, standard iterative algorithm is performed, running in parallel. 
 +
 +Note that explicit broadcasting used for ''centers'' object is not strictly needed -- if we used ''old_centers = centers'', the example would work too, but it would send a copy of ''old_centers'' to //every distributed task//, instead of once to every machine.
 +
 +For illustration, Scala version of the example follows. It works exactly as the Python version and uses ''breeze.linalg.Vector'' providing linear algebraic operations.
 +<file scala>
 +import breeze.linalg.Vector
 +
 +type Vector = breeze.linalg.Vector[Double]
 +type Vectors = Array[Vector]
 +
 +def closestPoint(point : Vector, centers : Vectors) : Double =
 +  centers.map(center => (center-point).norm(2)).zipWithIndex.min._2
 +
 +val lines = sc.textFile("/net/projects/hadoop/examples/inputs/points-small/points.txt")
 +val data = lines.map(line => Vector(line.split("\\s+").map(_.toDouble))).cache()
 +
 +val K = 50
 +val epsilon = 1e-3
 +
 +var i = 0
 +var centers_change = Double.PositiveInfinity
 +var centers = data.takeSample(false, K)
 +while (i < 10 && centers_change > epsilon) {
 +  val old_centers = sc.broadcast(centers)
 +  centers = (data
 +             // For each point, find its closes center index.
 +             .map(point => (closestPoint(point, old_centers.value), (point, 1)))
 +             // Sum points and counts in each cluster.
 +             .reduceByKey((a, b) => (a._1+b._1, a._2+b._2))
 +             // Sort by cluster index.
 +             .sortByKey()
 +             // Compute the new centers by averaging corresponding points.
 +             .map({case (index, (sum, count)) => sum :/ count.toDouble})
 +             .collect())
 +
 +  // Compute change in center positions.
 +  centers_change = (centers zip old_centers.value).map({case (a,b) => (a-b).norm(2)}).sum
 +  old_centers.unpersist()
 +  i += 1
 +}
 +
 +print(centers.deep)
 +</file>
  

[ Back to the navigation ] [ Back to the content ]