85 lines
3.0 KiB
C
85 lines
3.0 KiB
C
|
//
|
||
|
|
// PAA.h
|
||
|
|
// FastDTW-x
|
||
|
|
//
|
||
|
|
// Created by Melo Yao on 12/5/13.
|
||
|
|
// Copyright (c) 2013 melo.yao. All rights reserved.
|
||
|
|
//
|
||
|
|
|
||
|
|
#ifndef __FastDTW_x__PAA__
|
||
|
|
#define __FastDTW_x__PAA__
|
||
|
|
|
||
|
|
#include "Foundation.h"
|
||
|
|
#include <vector>
|
||
|
|
#include "TimeSeries.h"
|
||
|
|
#include "FDAssert.h"
|
||
|
|
#include <cmath>
|
||
|
|
FD_NS_START
|
||
|
|
template <typename ValueType,JInt nDimension>
|
||
|
|
class PAA : public TimeSeries<ValueType, nDimension>
|
||
|
|
{
|
||
|
|
vector<JInt> _aggPtSize;
|
||
|
|
JInt _originalLength;
|
||
|
|
public:
|
||
|
|
|
||
|
|
PAA(const TimeSeries<ValueType,nDimension>& ts, JInt shrunkSize):TimeSeries<ValueType, nDimension>(), _originalLength(ts.size()),_aggPtSize(shrunkSize)
|
||
|
|
{
|
||
|
|
FDASSERT(shrunkSize>0 && shrunkSize <= ts.size(),"ERROR: The size of an aggregate representation must be greater than zero and \nno larger than the original time series. (shrunkSize=%ld , origSize=%ld).",shrunkSize,ts.size());
|
||
|
|
// Ensures that the data structure storing the time series will not need
|
||
|
|
// to be expanded more than once. (not necessary, for optimization)
|
||
|
|
TimeSeries<ValueType,nDimension>::setMaxCapacity(shrunkSize);
|
||
|
|
TimeSeries<ValueType,nDimension>::setLabels(*ts.getLabels());
|
||
|
|
JDouble reducedPtSize = ts.size()/(JDouble)shrunkSize;
|
||
|
|
JInt ptToReadFrom(0);
|
||
|
|
JInt ptToReadTo;
|
||
|
|
while (ptToReadFrom < ts.size()) {
|
||
|
|
ptToReadTo = (JInt)round(reducedPtSize*(TimeSeries<ValueType,nDimension>::size()+1)) -1;
|
||
|
|
|
||
|
|
JInt ptsToRead = ptToReadTo - ptToReadFrom + 1;
|
||
|
|
JDouble timeSum(0.0);
|
||
|
|
ValueType measurementSums[nDimension];
|
||
|
|
fill(measurementSums, measurementSums+nDimension, 0);
|
||
|
|
for (JInt pt = ptToReadFrom; pt<=ptToReadTo; ++pt) {
|
||
|
|
const MeasurementVector<ValueType, nDimension> *currentPoint = ts.getMeasurementVector(pt);
|
||
|
|
timeSum += ts.getTimeAtNthPoint(pt);
|
||
|
|
for (JInt dim = 0; dim<ts.numOfDimensions(); ++dim) {
|
||
|
|
measurementSums[dim] += (*currentPoint)[dim];
|
||
|
|
}
|
||
|
|
}
|
||
|
|
// Determine the average value over the range ptToReadFrom...ptToReadFrom.
|
||
|
|
timeSum = timeSum / ptsToRead;
|
||
|
|
for (JInt dim = 0; dim<ts.numOfDimensions(); dim++) {
|
||
|
|
measurementSums[dim] = measurementSums[dim] / ptsToRead;
|
||
|
|
}
|
||
|
|
_aggPtSize[TimeSeries<ValueType,nDimension>::size()] = ptsToRead;
|
||
|
|
TimeSeries<ValueType,nDimension>::addLast(timeSum,TimeSeriesPoint<ValueType,nDimension>(measurementSums));
|
||
|
|
ptToReadFrom = ptToReadTo + 1;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
JInt originalSize() const
|
||
|
|
{
|
||
|
|
return _originalLength;
|
||
|
|
}
|
||
|
|
|
||
|
|
JInt aggregatePtSize(JInt ptIndex) const
|
||
|
|
{
|
||
|
|
return _aggPtSize[ptIndex];
|
||
|
|
}
|
||
|
|
|
||
|
|
void print(ostream& stream) const
|
||
|
|
{
|
||
|
|
TimeSeries<ValueType, nDimension>::print(stream);
|
||
|
|
stream<<"original len:"<<_originalLength<<"\n";
|
||
|
|
for(JInt i = 0;i<_aggPtSize.size();++i)
|
||
|
|
{
|
||
|
|
stream<<_aggPtSize[i] << ",";
|
||
|
|
}
|
||
|
|
stream<<"\n";
|
||
|
|
|
||
|
|
}
|
||
|
|
};
|
||
|
|
|
||
|
|
FD_NS_END
|
||
|
|
#endif /* defined(__FastDTW_x__PAA__) */
|