diff options
author | Matei Zaharia <matei@eecs.berkeley.edu> | 2010-03-29 16:17:55 -0700 |
---|---|---|
committer | Matei Zaharia <matei@eecs.berkeley.edu> | 2010-03-29 16:17:55 -0700 |
commit | df29d0ea4c8b7137fdd1844219c7d489e3b0d9c9 (patch) | |
tree | 3f925c0d109b789ce845762a9e09d24329749eb8 /src/examples/SparkLR.scala | |
download | spark-df29d0ea4c8b7137fdd1844219c7d489e3b0d9c9.tar.gz spark-df29d0ea4c8b7137fdd1844219c7d489e3b0d9c9.tar.bz2 spark-df29d0ea4c8b7137fdd1844219c7d489e3b0d9c9.zip |
Initial commit
Diffstat (limited to 'src/examples/SparkLR.scala')
-rw-r--r-- | src/examples/SparkLR.scala | 48 |
1 files changed, 48 insertions, 0 deletions
diff --git a/src/examples/SparkLR.scala b/src/examples/SparkLR.scala new file mode 100644 index 0000000000..34574f5640 --- /dev/null +++ b/src/examples/SparkLR.scala @@ -0,0 +1,48 @@ +import java.util.Random +import Vector._ +import spark._ + +object SparkLR { + val N = 10000 // Number of data points + val D = 10 // Numer of dimensions + val R = 0.7 // Scaling factor + val ITERATIONS = 5 + val rand = new Random(42) + + case class DataPoint(x: Vector, y: Double) + + def generateData = { + def generatePoint(i: Int) = { + val y = if(i % 2 == 0) -1 else 1 + val x = Vector(D, _ => rand.nextGaussian + y * R) + DataPoint(x, y) + } + Array.fromFunction(generatePoint _)(N) + } + + def main(args: Array[String]) { + if (args.length == 0) { + System.err.println("Usage: SparkLR <host> [<slices>]") + System.exit(1) + } + val sc = new SparkContext(args(0), "SparkLR") + val numSlices = if (args.length > 1) args(1).toInt else 2 + val data = generateData + + // Initialize w to a random value + var w = Vector(D, _ => 2 * rand.nextDouble - 1) + println("Initial w: " + w) + + for (i <- 1 to ITERATIONS) { + println("On iteration " + i) + val gradient = sc.accumulator(Vector.zeros(D)) + for (p <- sc.parallelize(data, numSlices)) { + val scale = (1 / (1 + Math.exp(-p.y * (w dot p.x))) - 1) * p.y + gradient += scale * p.x + } + w -= gradient.value + } + + println("Final w: " + w) + } +} |