/// <summary>
/// Use this model to evaluate a context and return an array of the likelihood of each outcome given that context.
/// </summary>
/// <param name="context">The names of the predicates which have been observed at the present decision point.</param>
/// <param name="values">This is where the distribution is stored.</param>
/// <param name="outsums">The outsums.</param>
/// <returns>The normalized probabilities for the outcomes given the context. The indexes of the double[] are the outcome ids, and the actual string representation of the outcomes can be obtained from the method getOutcome(int i).</returns>
public double[] Eval(string[] context, float[] values, double[] outsums)
{
var items = new int[context.Length];
for (int i = 0; i < context.Length; i++)
{
items[i] = map[context[i]];
}
prior.LogPrior(outsums, items, values);
return(Eval(items, values, outsums, evalParameters));
}
private void Compute(int threadIndex, int startIndex, int length)
{
var modelDistribution = new double[numOutcomes];
// to avoid unnecessary comparisons during the loop
var token = monitor != null
? monitor.Token
: new CancellationToken();
for (var ei = startIndex; ei < startIndex + length; ei++)
{
token.ThrowIfCancellationRequested();
if (values != null)
{
prior.LogPrior(modelDistribution, contexts[ei], values[ei]);
GISModel.Eval(contexts[ei], values[ei], modelDistribution, evalParams);
}
else
{
prior.LogPrior(modelDistribution, contexts[ei]);
GISModel.Eval(contexts[ei], modelDistribution, evalParams);
}
for (var j = 0; j < contexts[ei].Length; j++)
{
var pi = contexts[ei][j];
if (predicateCounts[pi] < cutoff)
{
continue;
}
var activeOutcomes = modelExpects[threadIndex][pi].Outcomes;
for (var aoi = 0; aoi < activeOutcomes.Length; aoi++)
{
var oi = activeOutcomes[aoi];
if (values != null && values[ei] != null)
{
modelExpects[threadIndex][pi].UpdateParameter(aoi, modelDistribution[oi] * values[ei][j] * numTimesEventsSeen[ei]);
}
else
{
modelExpects[threadIndex][pi].UpdateParameter(aoi, modelDistribution[oi] * numTimesEventsSeen[ei]);
}
}
}
threadLoglikelihood[threadIndex] += Math.Log(modelDistribution[outcomeList[ei]]) * numTimesEventsSeen[ei];
threadNumEvents[threadIndex] += numTimesEventsSeen[ei];
#if !DEBUG
// only when we have a monitor to report...
if (monitor != null)
{
#endif
var max = 0;
for (var oi = 0; oi < numOutcomes; oi++)
{
if (modelDistribution[oi] > modelDistribution[max])
{
max = oi;
}
}
if (max == outcomeList[ei])
{
threadNumCorrect[threadIndex] += numTimesEventsSeen[ei];
}
#if !DEBUG
}
#endif
}
}
