/**
* Get the top values from a list of value-frequencies
*
* @param values
* @param bias
* @return
*/
public List<ValueFreq<T>> getTopValueUsingPriorityQueue(List<ValueFreq<T>> valueFreqs, int bias) {
// sort by values
PriorityQueue<ValueFreq<T>> pq =
new PriorityQueue<ValueFreq<T>>(
bias,
new Comparator<ValueFreq<T>>() {
public int compare(ValueFreq bv1, ValueFreq bv2) {
if ((double) (Double) bv1.getValue() < (double) (Double) bv2.getValue()) return -1;
else if ((double) (Double) bv1.getValue() == (double) (Double) bv2.getValue())
return 0;
else return 1;
}
});
// maintain a priority queue
for (int i = 0; i < valueFreqs.size(); i++) {
if (i < bias) {
pq.add(valueFreqs.get(i));
} else {
pq.add(valueFreqs.get(i));
pq.poll();
}
}
LinkedList<ValueFreq<T>> biasValues = new LinkedList<ValueFreq<T>>();
Iterator it = pq.iterator();
while (it.hasNext()) {
ValueFreq<T> bv = (ValueFreq<T>) it.next();
biasValues.add(new ValueFreq<T>(bv.getValue(), bv.getFreq()));
}
return biasValues;
}
/**
* Return all values
*
* @param vf
* @return
*/
public List<T> getValues(List<ValueFreq<T>> vf) {
ArrayList<T> list = new ArrayList<T>();
Iterator it = vf.iterator();
while (it.hasNext()) {
ValueFreq<T> bv = (ValueFreq) it.next();
for (int i = 0; i < bv.getFreq(); i++) list.add(bv.getValue());
}
return list;
}
/**
* Calculate a value list
*
* @param vf
* @return
*/
public List<T> getDistinctValues(List<ValueFreq<T>> vf) {
ArrayList<T> list = new ArrayList<T>();
Iterator it = vf.iterator();
while (it.hasNext()) {
ValueFreq<T> bv = (ValueFreq) it.next();
list.add(bv.getValue());
}
return list;
}
/**
* Calculate a frequency list
*
* @param vf
* @return
*/
public List<Integer> getFreqs(List<ValueFreq<T>> vf) {
ArrayList<Integer> list = new ArrayList<Integer>();
Iterator it = vf.iterator();
while (it.hasNext()) {
ValueFreq bv = (ValueFreq) it.next();
list.add(bv.getFreq());
}
return list;
}
/**
* Get sum of all freqs
*
* @param valueFreqs
* @return
*/
public int totalNumbers(List<ValueFreq<T>> valueFreqs) {
int totalNumber = 0;
Iterator it = valueFreqs.iterator();
while (it.hasNext()) {
ValueFreq vf = (ValueFreq) it.next();
totalNumber += vf.getFreq();
}
return totalNumber;
}
public List<Integer> getFreqs(List<ValueFreq<T>> valueFreqs, int start, int end) {
List<ValueFreq<T>> tmp_valueFreqs = valueFreqs.subList(start, end);
ArrayList<Integer> list = new ArrayList<Integer>();
Iterator it = tmp_valueFreqs.iterator();
while (it.hasNext()) {
ValueFreq<T> bv = (ValueFreq<T>) it.next();
list.add(bv.getFreq());
}
return list;
}
/**
* Return weight mean, weight is frequency
* http://en.wikipedia.org/wiki/Weighted_variance#Weighted_sample_variance
*
* @param list
* @return
*/
public double weightedMean(List<ValueFreq> list) {
double wm = 0.0;
double sumOfWeight = 0.0;
Iterator it = list.iterator();
while (it.hasNext()) {
ValueFreq vr = (ValueFreq) it.next();
double weight = vr.getFreq();
sumOfWeight += weight;
wm += Double.parseDouble(vr.getValue().toString()) * weight;
}
return wm / sumOfWeight;
}
/**
* Calculate weighted variance
*
* @param list
* @return
*/
public double weightedVarianceSample(List<ValueFreq> list) {
// weighted mean
double wm = weightedMean(list);
// calculate sum of weight
double sumOfWeight = 0.0;
// weighted variance
double variance = 0.0;
Iterator it = list.iterator();
while (it.hasNext()) {
ValueFreq vr = (ValueFreq) it.next();
double weight = vr.getFreq();
sumOfWeight += weight;
variance += weight * Math.pow(Double.parseDouble(vr.getValue().toString()) - wm, 2);
}
return variance / (sumOfWeight - 1);
}
/**
* Total area of a distribution
*
* @param valueFreqs
* @return
*/
public double area(List<ValueFreq<T>> valueFreqs) {
double totalArea = 0.0;
Iterator it = valueFreqs.iterator();
T lastValue = null;
int lastFreq = 0;
if (it.hasNext()) {
ValueFreq<T> vf = (ValueFreq<T>) it.next();
lastValue = vf.getValue();
lastFreq = vf.getFreq();
}
while (it.hasNext()) {
ValueFreq<T> vf = (ValueFreq<T>) it.next();
totalArea +=
0.5
* (vf.getFreq() + lastFreq)
* ((Double) vf.getValue() - Double.valueOf(lastValue.toString()));
lastValue = vf.getValue();
lastFreq = vf.getFreq();
}
return totalArea;
}
public List<Double> areas(List<ValueFreq<T>> valueFreqs) {
List<Double> areas = new ArrayList<Double>();
Iterator it = valueFreqs.iterator();
T lastValue = null;
int lastFreq = 0;
if (it.hasNext()) {
ValueFreq<T> vf = (ValueFreq<T>) it.next();
lastValue = vf.getValue();
lastFreq = vf.getFreq();
}
while (it.hasNext()) {
ValueFreq<T> vf = (ValueFreq<T>) it.next();
double area =
0.5
* (vf.getFreq() + lastFreq)
* ((Double) vf.getValue() - Double.valueOf(lastValue.toString()));
areas.add(area);
lastValue = vf.getValue();
lastFreq = vf.getFreq();
}
return areas;
}