Class BTable

src/com/winvector/variables/BTable.java:29–307

very low dimensional re-encoding of levels @author johnmount

 *
 */
public final class BTable {
	public VariableEncodings oldAdapter;
	public VariableEncodings newAdapter;
	public double[] warmStart = null;
	public Iterable<BurstMap> sample = null;
	
	private static final class BSampler implements SerialObserver<BurstMap> {
		public final VariableEncodings oldAdapter;
		public final ResevoirSampler<BurstMap> sampler;

		public BSampler(final VariableEncodings oldAdapter, final long randSeed) {
			this.oldAdapter = oldAdapter;
			sampler = new ResevoirSampler<BurstMap>(100000,randSeed);
		}
		
		@Override
		public void observe(final BurstMap row) {
			final double weight = oldAdapter.weight(row);
			if(weight>0.0) {
				final String resStr = row.getAsString(oldAdapter.def().resultColumn);
				final int category = oldAdapter.category(resStr);
				if(category>=0) {
					sampler.observe(row);
				}
			}
		}
	}
	
	

	private static final class GeneralIndicator {		
		@SuppressWarnings("unused")
		public final String var;
		public final String name;
		public final Map<String,Double> levelEncodings = new HashMap<String,Double>(); // indexed by levels of var
		public final int warmStartOutcome;
		
		public GeneralIndicator(final String var, final String name, final int nlevels, final int warmStartOutcome) {
			this.var = var;
			this.name = name;
			this.warmStartOutcome = warmStartOutcome;
		}
				
		public double normSq() {
			double nsq = 0.0;
			for(final double v: levelEncodings.values()) {
				nsq += v*v;
			}
			return nsq;
		}
	}
	

	public static ArrayList<GeneralIndicator> encode(final BStat stat, final String variable, final VariableEncodings oldAdapter, final double[] oldX) {
		final ArrayList<GeneralIndicator> res = new ArrayList<GeneralIndicator>();
		final double smooth = 0.5;
		final double sumAll = stat.sumTotal + smooth;
		final Set<String> levels = oldAdapter.def().catLevels.get(variable).keySet();