// not occur immediately after the last training event. Since the time
// index is artificial, all we can do for future predictions is assume
// that the n instances provided to the primeForecaster() method overlap
// the last n instances of the training data and that future predictions
// occur from the last known artificial time value + 1.
m_artificialTimeMaker = new AddID();
m_artificialTimeMaker.setAttributeName("ArtificialTimeIndex");
m_artificialTimeMaker.setIDIndex("last");
m_artificialTimeMaker.setInputFormat(result);
result = Filter.useFilter(result, m_artificialTimeMaker);
m_useArtificialTimeIndex = true;