//private bool AllVarianceZero(Dictionary<string, GaussianStatistics> caseNameToTarget)
private static bool AllVarianceZero(IEnumerable <Leaf> LeafCollection, Converter <Leaf, SufficientStatistics> caseNameToTarget)
{
bool varianceIsZero = false; //If empty input, then return false
bool firstTime = true;
foreach (Leaf leaf in LeafCollection)
{
SufficientStatistics stats = caseNameToTarget(leaf);
if (stats.IsMissing())
{
continue;
}
GaussianStatistics gaussianStatistics = (GaussianStatistics)stats;
SpecialFunctions.CheckCondition((gaussianStatistics.Variance == 0) == (gaussianStatistics.SampleSize == 1), "Variance must be zero exactly when the sample size is 1");
if (firstTime)
{
firstTime = false;
varianceIsZero = (gaussianStatistics.Variance == 0);
}
else
{
SpecialFunctions.CheckCondition(varianceIsZero == (gaussianStatistics.Variance == 0), "If any variances are zero, then all must be zero");
}
}
return(varianceIsZero);
}
internal static bool Parse(string val, out SufficientStatistics result)
{
result = null;
if (val.Equals("null", StringComparison.InvariantCultureIgnoreCase))
{
result = GetMissingInstance();
return(false);
}
else
{
string[] fields = val.Split(',');
if (!(fields.Length == 3))
{
return(false);
}
double mean, variance;
int sampleSize;
if (double.TryParse(fields[0], out mean) &&
double.TryParse(fields[1], out variance) &&
int.TryParse(fields[2], out sampleSize))
{
result = GaussianStatistics.GetInstance(mean, variance, sampleSize);
return(true);
}
return(false);
}
}
public static Converter <Leaf, SufficientStatistics> DictionaryToLeafMap(Dictionary <string, GaussianStatistics> caseIdToNonMissingValue)
{
return(delegate(Leaf leaf)
{
string name = leaf.CaseName;
if (name == null || !caseIdToNonMissingValue.ContainsKey(name))
{
return GaussianStatistics.GetMissingInstance();
}
else
{
return caseIdToNonMissingValue[name];
}
});
}
public override bool Equals(object obj)
{
if (IsMissing() && obj is SufficientStatistics && ((SufficientStatistics)obj).IsMissing())
{
return(true);
}
GaussianStatistics other = obj as GaussianStatistics;
if (other == null)
{
return(false);
}
return(_mean == other._mean && _variance == other._variance && _sampleSize == other._sampleSize);
}
public static SufficientStatistics Parse(string val)
{
SufficientStatistics result;
if (
MissingStatistics.Parse(val, out result) ||
GaussianStatistics.Parse(val, out result) ||
BooleanStatistics.Parse(val, out result) ||
DiscreteStatistics.Parse(val, out result) ||
ContinuousStatistics.Parse(val, out result))
{
return(result);
}
throw new ArgumentException(string.Format("Unable to parse {0} into an instance of ISufficientStatistics", val));
}
//public override Converter<Leaf, SufficientStatistics> CreateTargetSufficientStatisticsMap(Dictionary<string, ISufficientStatistics> caseIdToNonMissingValue)
//{
// return ISufficientStatistics.DictionaryToLeafMap(caseIdToNonMissingValue);
//}
public override Converter <Leaf, SufficientStatistics> CreateAlternativeSufficientStatisticsMap(
Converter <Leaf, SufficientStatistics> predictorDistributionClassFunction,
Converter <Leaf, SufficientStatistics> targetDistributionClassFunction)
{
return(delegate(Leaf leaf)
{
SufficientStatistics predStats = predictorDistributionClassFunction(leaf);
SufficientStatistics targetStats = targetDistributionClassFunction(leaf);
// bail on missing data.
if (predStats.IsMissing() || targetStats.IsMissing())
{
return GaussianStatistics.GetMissingInstance();
}
else
{
return targetStats;
}
});
}
private MessageGaussian Initalize(ConditionalGaussianDistributionParams dist, Leaf leaf, OptimizationParameterList gaussianParameters)
{
Debug.Assert(!LeafToTargetStatistics(leaf).IsMissing());
GaussianStatistics gaussianStatistics = (GaussianStatistics)LeafToTargetStatistics(leaf);
SpecialFunctions.CheckCondition(gaussianStatistics != null, "why is caseNameToTargetOrNull unknown?");
if (gaussianStatistics.SampleSize == 1)
{
double z = gaussianStatistics.Mean;
double logK = -Math.Log(dist.LinearCoefficient);
Debug.Assert(!double.IsNaN(logK)); // real assert - in release mode, the NaN is OK.
double a = (z - dist.Mean) / dist.LinearCoefficient;
double v = dist.Variance / Math.Pow(dist.LinearCoefficient, 2);
MessageGaussian message = MessageGaussian.GetInstance(logK, a, v);
return(message);
}
else
{
double vNoise = GaussianDistribution.GetSamplingVariance(gaussianParameters);
double logKMult = (
-Math.Log(gaussianStatistics.SampleSize)
- gaussianStatistics.SampleSize
* (gaussianStatistics.Variance + vNoise * Log2PI + vNoise * Math.Log(vNoise))
/ vNoise
+ Math.Log(2 * Math.PI * vNoise)
) / 2.0;
double aMult = gaussianStatistics.Mean;
double vMult = vNoise / (double)gaussianStatistics.SampleSize;
double logK = logKMult - Math.Log(dist.LinearCoefficient);
double a = (aMult - dist.Mean) / dist.LinearCoefficient;
double v = (vMult + dist.Variance) / Math.Pow(dist.LinearCoefficient, 2);
MessageGaussian message = MessageGaussian.GetInstance(logK, a, v);
return(message);
}
}
