/**
* Measure of statistical dependence between two variables
*
* <p>http://en.wikipedia.org/wiki/Spearman's_rank_correlation_coefficient
*
* @param X
* @param Y
* @return
*/
public double spearmanCC(List<T> X, List<T> Y) {
LinkedList<ValueRank> valueRankX = new LinkedList<ValueRank>();
LinkedList<ValueRank> valueRankY = new LinkedList<ValueRank>();
Iterator itX = X.iterator();
Iterator itY = Y.iterator();
int size = 0;
while (itX.hasNext() && itY.hasNext()) {
size++;
valueRankX.add(new ValueRank<T>((T) itX.next(), size));
valueRankY.add(new ValueRank<T>((T) itY.next(), size));
}
Collections.sort(valueRankX);
Collections.sort(valueRankY);
itX = valueRankX.iterator();
itY = valueRankY.iterator();
LinkedList<Double> rankX = new LinkedList<Double>();
LinkedList<Double> rankY = new LinkedList<Double>();
while (itX.hasNext() && itY.hasNext()) {
ValueRank vrx = (ValueRank) itX.next();
ValueRank vry = (ValueRank) itY.next();
rankX.add(vrx.getRank());
rankY.add(vry.getRank());
}
Stats<Double> stat = new Stats<Double>();
return stat.personCC(rankX, rankY);
}
/**
* Get the probabilities from a list of frequencies
*
* @param freqs
* @return
*/
public List<Double> probsFromFreqs(List<Integer> freqs) {
ArrayList<Double> probs = new ArrayList<Double>();
Stats<Integer> stat = new Stats<Integer>();
double sum = stat.sum(freqs);
Iterator it = freqs.iterator();
while (it.hasNext()) {
double p = Double.parseDouble(it.next().toString()) / sum;
probs.add(p);
}
return probs;
}
