- net.ClearAllEvidence();
net.SliceCount = myTable.Rows.Count;
foreach (DataRow myRow in myTable.Rows)- setEvidenceProcess(myRow, actionNode, net);
public void setEvidenceProcess(DataRow row, AppNode actionNode, Network net)
{- ...
foreach (DictionaryEntry de in slStateLabels)
{
...- {
- net.SetTemporalEvidence(nHandler, IDtemp, outcomeIndex); //(int nodeHandler, int slice, int outComeIndex)
- {
...
net.UpdateBeliefs();
...
int nHandle = node.getNodeID();
...
double[] probArray = net.GetNodeValue(nHandle);
We did the same thing in GeNie, and compared node by node and found the probabilities generated are different.
Could you tell us why and do you have new API for DBN?