HopfieldNetworkTest Class Reference

Test of HopfieldNetwork Class. More...

Inheritance diagram for HopfieldNetworkTest:

Collaboration diagram for HopfieldNetworkTest:

Protected Member Functions
vector< OverlapResolverNodeInfo >	getInput ()
	Create sample for test. More...

Protected Attributes
vector< OverlapResolverNodeInfo >	m_trackCandidateInfos
	Container on which the Hopfield Algorithm runs. More...
vector< OverlapResolverNodeInfo >	m_qiTrackOverlap
	Container on which the Greedy Algorithm (Scrooge) runs. More...
unsigned int	myTrueTracks = 10
	Number of true tracks.

Detailed Description

Test of HopfieldNetwork Class.

Definition at line 25 of file hopfieldNetwork.cc.

Member Function Documentation

getInput()

vector<OverlapResolverNodeInfo> getInput ( )

inlineprotected

Create sample for test.

Returns

test sample info

Definition at line 43 of file hopfieldNetwork.cc.

  {
    vector<OverlapResolverNodeInfo> trackCandidateInfos;
    unsigned int myNTrackCands  = 200;
    unsigned int diff           = myNTrackCands - myTrueTracks;
    unsigned int nOverlaps      = 6;
 
    //Create competitor IDs
    vector <vector <unsigned short> > competitorIDMatrix(myNTrackCands);
    for (unsigned int ii = 0; ii < myNTrackCands; ii++) {
      for (unsigned int jj = 0; jj < nOverlaps; jj++) {
        unsigned short bkgIndex = (rand() % diff) + myTrueTracks;
        competitorIDMatrix[ii].push_back(bkgIndex);
        competitorIDMatrix[bkgIndex].push_back(ii);
      }
    }
 
    //Make the actual OverlapResolverNodeInfos
    for (unsigned int ii = 0; ii < myNTrackCands; ii++) {
      float qualityIndicator = 0;
      if (ii < myTrueTracks) {
        qualityIndicator = static_cast<float>(rand() % 100) / 100.;
        B2INFO("Track Index" << ii << ", TrueQI: " << qualityIndicator);
      } else {
        qualityIndicator = 1 / (static_cast<float>(rand() % 100) + 1.2);
        B2INFO("Track Index" << ii << ", FakeQI: " << qualityIndicator);
      }
      trackCandidateInfos.emplace_back(qualityIndicator, ii, competitorIDMatrix[ii], 0.8);
    }
    return trackCandidateInfos;
  }

Member Data Documentation

m_qiTrackOverlap

vector<OverlapResolverNodeInfo> m_qiTrackOverlap

protected

Container on which the Greedy Algorithm (Scrooge) runs.

Output comes in form of qiTrackOverlap.isActive, which is simply a bool, if the corresponding track was taken or not.

Definition at line 36 of file hopfieldNetwork.cc.

m_trackCandidateInfos

vector<OverlapResolverNodeInfo> m_trackCandidateInfos

protected

Container on which the Hopfield Algorithm runs.

Output comes in form of tcInfo4Hopfield.neuronValue. Typically > 0.7 signal, < 0.7 bkg;

Definition at line 31 of file hopfieldNetwork.cc.

---

The documentation for this class was generated from the following file:

• tracking/trackFindingVXD/trackSetEvaluator/tests/hopfieldNetwork.cc