tdunning

package com.mapr.bench;

import com.google.caliper.Benchmark;
import com.google.caliper.runner.Running;
import org.apache.commons.math3.util.FastMath;
import org.junit.Assert;
import org.junit.Test;

public class Distance {
    @Benchmark

tdunning

# for auc()
> library(pROC)

# for performance plots
> library(ROCR)
Loading required package: gplots
KernSmooth 2.23 loaded
Copyright M. P. Wand 1997-2009

Attaching package: ‘gplots’

tdunning

  public static class BigDecimalWritable implements Writable {
    private BigDecimal value;

    public BigDecimalWritable(BigDecimal value) {
      this.value = value;
    }
    
    public BigDecimal value() {
      return value;
    }

tdunning

  @Test
  public void testStats() {
    // the reference limits here were derived using a numerical simulation where I took
    // 10,000 samples from the distribution in question and computed the stats from that
    // sample to get min, 25%-ile, median and so on.  I did this 1000 times to get 5% and
    // 95% confidence limits for those values.

    // symmetrical, well behaved
    System.out.printf("normal\n");
    check(normal(10000));

tdunning

  Set<String> common = Sets.newHashSet(firstListOfEmails);
  common.retainAll(secondListOfEmails);

tdunning

public class HbaseLookup {
    static final org.slf4j.Logger logger = org.slf4j.LoggerFactory.getLogger(TrigoMathFunctions.class);

    private HbaseLookup(){}

    @FunctionTemplate(name = "hLookup", scope = FunctionScope.SIMPLE, nulls = NullHandling.NULL_IF_NULL)
    public static class Lookup implements DrillSimpleFunc {

        @Param VarCharHolder table;                // the table to read from

tdunning

# picking the corners of the hyper cube at random usually gives us a good selection
d = 0
while (d == 0) {
    centers = matrix(runif(10*10)>0.5, ncol=10) + 0
    # but occasionally we get a duplicate row that is easily detected
    d = det(centers)
}

# start x out by selecting clusters
x = data.frame(n = ceiling(runif(10000,1e-10,10)))

tdunning

# Experiments with t-digest in R

standard.size.bound = function(n, q) {
    4 * n * q * (1-q)
}

constant.size.bound = function(n, q) {
    n
}

tdunning

Log in to the cluster:

    ted:downloads$ ssh se-node10.se.lab
    Last login: Mon Mar 23 17:35:37 2015 from 10.250.0.220
    Please check the cluster reservation calendar:
    https://www.google.com/calendar/embed?src=maprtech.com_2d38343133383836382d313737%40resource.calendar.google.com

Poke around looking for my volume and such:

    [tdunning@se-node10 ~]$ ls /mapr/se1/user/t

tdunning

import fileinput
from string import join
import json
import csv
import json
### read the output from MAHOUT and collect into hash ###
with open('x','rb') as csv_file:
    csv_reader = csv.reader(csv_file,delimiter='\t')
    old_id = ""
    indicators = []