public List<TrainAndTestReportCrisp> computeAvgCTSperM(List<TrainAndTestReportCrisp> reportsCTS) {
weightsCrisp.clear();
weightsInterval.clear();
Map<Model, List<TrainAndTestReportCrisp>> mapForAvg = new HashMap<>();
for (TrainAndTestReportCrisp r : reportsCTS) {
if (mapForAvg.containsKey(r.getModel())) {
mapForAvg.get(r.getModel()).add(r);
} else {
List<TrainAndTestReportCrisp> l = new ArrayList<>();
l.add(r);
mapForAvg.put(r.getModel(), l);
}
}
List<TrainAndTestReportCrisp> avgReports = new ArrayList<>();
for (Model model : mapForAvg.keySet()) {
List<TrainAndTestReportCrisp> l = mapForAvg.get(model);
if (l.size() == 1) { // does not make sense to compute average over one series
// do not compute anything
} else {
TrainAndTestReportCrisp thisAvgReport = computeAvgCTS(l, model);
if (thisAvgReport != null) {
avgReports.add(thisAvgReport);
} else { // should never happen for the same method
System.err.println("nerovnake percenttrain v ramci 1 modelu pri avg CTS per method :/");
}
}
}
return avgReports;
}
private static void recomputeTrainTestSets(TrainAndTestReport r) { // TODO a nicer way? :)
if (r instanceof TrainAndTestReportCrisp) {
TrainAndTestReportCrisp rep = (TrainAndTestReportCrisp) r;
double[] newFittedValues =
Arrays.copyOfRange(rep.getFittedValues(), 0, rep.getNumTrainingEntries());
double[] newForecastValsTest =
Arrays.copyOfRange(
rep.getFittedValues(), rep.getNumTrainingEntries(), rep.getFittedValues().length);
rep.setFittedValues(newFittedValues);
rep.setForecastValuesTest(newForecastValsTest);
double[] newRealTrain =
Arrays.copyOfRange(rep.getRealOutputsTrain(), 0, rep.getNumTrainingEntries());
double[] newRealTest =
Arrays.copyOfRange(
rep.getRealOutputsTrain(),
rep.getNumTrainingEntries(),
rep.getRealOutputsTrain().length);
rep.setRealOutputsTrain(newRealTrain);
rep.setRealOutputsTest(newRealTest);
rep.setErrorMeasures(
ErrorMeasuresUtils.computeAllErrorMeasuresCrisp(
Utils.arrayToList(newRealTrain),
Utils.arrayToList(newRealTest),
Utils.arrayToList(newFittedValues),
Utils.arrayToList(newForecastValsTest),
0)); // TODO I hope the 0 is not a problem
} else if (r instanceof TrainAndTestReportInterval) {
TrainAndTestReportInterval rep = (TrainAndTestReportInterval) r;
List<Interval> newFittedValues =
new ArrayList<>(rep.getFittedValues().subList(0, rep.getNumTrainingEntries()));
List<Interval> newForecastValsTest =
new ArrayList<>(
rep.getFittedValues()
.subList(rep.getNumTrainingEntries(), rep.getFittedValues().size()));
rep.setFittedValues(newFittedValues);
rep.setForecastValuesTest(newForecastValsTest);
List<Interval> realVals =
Utils.zipLowerUpperToIntervals(rep.getRealValuesLowers(), rep.getRealValuesUppers());
List<Interval> realValsTrain =
new ArrayList<>(realVals.subList(0, rep.getNumTrainingEntries()));
List<Interval> realValsTest =
new ArrayList<>(realVals.subList(rep.getNumTrainingEntries(), realVals.size()));
// TODO somehow add the actual distance and seasonality from params
rep.setErrorMeasures(
ErrorMeasuresUtils.computeAllErrorMeasuresInterval(
realValsTrain,
realValsTest,
newFittedValues,
newForecastValsTest,
new WeightedEuclideanDistance(0.5),
0));
}
}
public TrainAndTestReportCrisp computeAvgCTS(
List<TrainAndTestReportCrisp> reportsCTS, Model model) {
if (reportsCTS.size() == 1) { // does not make sense to compute average over one series
return reportsCTS.get(
0); // TODO do not return anything (but take care of it on the receiving end), because
// otherwise it draws twice. but then problems with "drawOnlyAVG"
} else {
if (!allTheSamePercentTrain(reportsCTS)) {
return null;
} else {
StringBuilder fittedValsAvgAll = new StringBuilder("(");
StringBuilder forecastValsTestAvgAll = new StringBuilder("(");
StringBuilder forecastValsFutureAvgAll = new StringBuilder("(");
StringBuilder sumWeightsTrain = new StringBuilder("(");
StringBuilder sumWeightsTest = new StringBuilder("(");
StringBuilder sumWeightsFuture = new StringBuilder("(");
boolean next = false;
for (TrainAndTestReportCrisp r : reportsCTS) {
if (next) {
fittedValsAvgAll.append(" + ");
forecastValsTestAvgAll.append(" + ");
forecastValsFutureAvgAll.append(" + ");
sumWeightsTrain.append(" + ");
sumWeightsTest.append(" + ");
sumWeightsFuture.append(" + ");
} else {
next = true;
}
double weightTrain = getWeightForModelTrain(r);
double weightTest = getWeightForModelTest(r);
double weightFuture = getWeightForModelFuture(r);
weightsCrisp.put(r.toString(), weightFuture);
sumWeightsTrain.append(weightTrain);
sumWeightsTest.append(weightTest);
fittedValsAvgAll
.append(weightTrain)
.append("*")
.append(Utils.arrayToRVectorString(r.getFittedValues()));
forecastValsTestAvgAll
.append(weightTest)
.append("*")
.append(Utils.arrayToRVectorString(r.getForecastValuesTest()));
forecastValsFutureAvgAll.append(weightFuture).append("*");
if (r.getForecastValuesFuture().length > 0) {
forecastValsFutureAvgAll.append(
Utils.arrayToRVectorString(r.getForecastValuesFuture()));
sumWeightsFuture.append(weightFuture);
} else {
forecastValsFutureAvgAll.append("0");
sumWeightsFuture.append("0");
}
}
sumWeightsTrain.append(")");
sumWeightsTest.append(")");
sumWeightsFuture.append(")");
fittedValsAvgAll.append(")/").append(sumWeightsTrain);
forecastValsTestAvgAll.append(")/").append(sumWeightsTest);
forecastValsFutureAvgAll.append(")/").append(sumWeightsFuture);
String avgAll =
"c("
+ fittedValsAvgAll
+ ","
+ forecastValsTestAvgAll
+ ","
+ forecastValsFutureAvgAll
+ ")";
MyRengine rengine = MyRengine.getRengine();
// and create a new report for this avg and add it to reportsCTS:
TrainAndTestReportCrisp thisAvgReport =
new TrainAndTestReportCrisp(model, "(" + getName() + ")", true);
double[] fittedValsAvg = rengine.evalAndReturnArray(fittedValsAvgAll.toString());
double[] forecastValsTestAvg =
rengine.evalAndReturnArray(forecastValsTestAvgAll.toString());
ErrorMeasuresCrisp errorMeasures =
ErrorMeasuresUtils.computeAllErrorMeasuresCrisp(
Utils.arrayToList(reportsCTS.get(0).getRealOutputsTrain()),
Utils.arrayToList(reportsCTS.get(0).getRealOutputsTest()),
Utils.arrayToList(fittedValsAvg),
Utils.arrayToList(forecastValsTestAvg),
0);
thisAvgReport.setErrorMeasures(errorMeasures);
double[] forecastValsFutureAvg =
rengine.evalAndReturnArray(forecastValsFutureAvgAll.toString());
thisAvgReport.setForecastValuesFuture(forecastValsFutureAvg);
thisAvgReport.setPlotCode("plot.ts(" + avgAll + ", lty=2)");
thisAvgReport.setFittedValues(fittedValsAvg);
thisAvgReport.setForecastValuesTest(forecastValsTestAvg);
thisAvgReport.setNumTrainingEntries(fittedValsAvg.length);
thisAvgReport.setRealOutputsTrain(reportsCTS.get(0).getRealOutputsTrain());
thisAvgReport.setRealOutputsTest(reportsCTS.get(0).getRealOutputsTest());
return thisAvgReport;
}
}
}