calculateStats.java

private static Pair<List<Integer>,List<Integer>> calculateStats(List<Integer> list)
{
	double[] data = new double[list.size()];
	for(int i = 0; i < list.size(); i++)
		data[i] = list.get(i);
	DescriptiveStatistics dStats = new DescriptiveStatistics(data);
	
	List<Integer> summary = new ArrayList<Integer>(5);
	summary.add((int) dStats.getMin()); //Minimum
	summary.add((int) dStats.getPercentile(25)); //Lower Quartile (Q1)
	summary.add((int) dStats.getPercentile(50)); //Middle Quartile (Median - Q2)
	summary.add((int) dStats.getPercentile(75)); //High Quartile (Q3)
	summary.add((int) dStats.getMax()); //Maxiumum
	
	List<Integer> outliers = new ArrayList<Integer>();
	if(list.size() > 5 && dStats.getStandardDeviation() > 0) //Only remove outliers if relatively normal
	{
		double mean = dStats.getMean();
		double stDev = dStats.getStandardDeviation();
		NormalDistribution normalDistribution = new NormalDistribution(mean, stDev);
		
		Iterator<Integer> listIterator = list.iterator();
		double significanceLevel = .50 / list.size();
		while(listIterator.hasNext())
		{
			int num = listIterator.next();
			double pValue = normalDistribution.cumulativeProbability(num);
			if(pValue < significanceLevel) //Chauvenet's Criterion for Outliers
			{
				outliers.add(num);
				listIterator.remove();
			}
		}
		
		if(list.size() != dStats.getN()) //If and only if outliers have been removed
		{
			double[] significantData = new double[list.size()];
			for(int i = 0; i < list.size(); i++)
				significantData[i] = list.get(i);
			dStats = new DescriptiveStatistics(significantData);
			summary.set(0, (int) dStats.getMin());
			summary.set(4, (int) dStats.getMax());
		}
	}
	
	return new Pair<List<Integer>,List<Integer>>(summary, outliers);
}