/// <summary>VMP message to <c>item</c>.</summary>
/// <param name="array">Incoming message from <c>array</c>. Must be a proper distribution. If all elements are uniform, the result will be uniform.</param>
/// <param name="index1">Constant value for <c>index1</c>.</param>
/// <param name="index2">Constant value for <c>index2</c>.</param>
/// <param name="result">Modified to contain the outgoing message.</param>
/// <returns>
/// <paramref name="result" />
/// </returns>
/// <remarks>
/// <para>The outgoing message is a distribution matching the moments of <c>item</c> as the random arguments are varied. The formula is <c>proj[sum_(array) p(array) factor(item,array,index1,index2)]</c>.</para>
/// </remarks>
/// <exception cref="ImproperMessageException">
/// <paramref name="array" /> is not a proper distribution.</exception>
/// <typeparam name="Distribution">The type of the distribution over an item.</typeparam>
public static Distribution ItemAverageLogarithm <Distribution>(
[SkipIfAllUniform] IArray2D <Distribution> array, int index1, int index2, Distribution result)
where Distribution : SettableTo <Distribution>
{
result.SetTo(array[index1, index2]);
return(result);
}
/// <summary>
/// Print the means of a 2-D array of Gaussians to the console
/// </summary>
/// <param name="matrix"></param>
private static void printMatrixToConsole(IArray2D <Gaussian> matrix)
{
for (int i = 0; i < matrix.GetLength(0); i++)
{
for (int j = 0; j < matrix.GetLength(1); j++)
{
Console.Write("{0,5:0.00}\t", matrix[i, j].GetMean());
}
Console.WriteLine("");
}
}
/// <summary>
/// Mean absolute row means
/// </summary>
/// <param name="matrix"></param>
/// <returns></returns>
private static double[] meanAbsoluteRowMeans(IArray2D <Gaussian> matrix)
{
double[] mam = new double[matrix.GetLength(0)];
double mult = 1.0 / ((double)matrix.GetLength(1));
for (int i = 0; i < matrix.GetLength(0); i++)
{
double sum = 0.0;
for (int j = 0; j < matrix.GetLength(1); j++)
{
sum += System.Math.Abs(matrix[i, j].GetMean());
}
mam[i] = mult * sum;
}
return(mam);
}
/// <summary>Evidence message for EP.</summary>
/// <param name="item">Incoming message from <c>item</c>.</param>
/// <param name="array">Incoming message from <c>array</c>.</param>
/// <param name="index1">Constant value for <c>index1</c>.</param>
/// <param name="index2">Constant value for <c>index2</c>.</param>
/// <returns>Logarithm of the factor's contribution the EP model evidence.</returns>
/// <remarks>
/// <para>The formula for the result is <c>log(sum_(item,array) p(item,array) factor(item,array,index1,index2) / sum_item p(item) messageTo(item))</c>. Adding up these values across all factors and variables gives the log-evidence estimate for EP.</para>
/// </remarks>
/// <typeparam name="Distribution">The type of the distribution over an item.</typeparam>
public static double LogEvidenceRatio <Distribution>(T item, IArray2D <Distribution> array, int index1, int index2)
where Distribution : CanGetLogProb <T>
{
return(LogAverageFactor(item, array, index1, index2));
}
/// <summary>Evidence message for EP.</summary>
/// <param name="item">Incoming message from <c>item</c>.</param>
/// <param name="array">Incoming message from <c>array</c>.</param>
/// <param name="index1">Constant value for <c>index1</c>.</param>
/// <param name="index2">Constant value for <c>index2</c>.</param>
/// <returns>Logarithm of the factor's average value across the given argument distributions.</returns>
/// <remarks>
/// <para>The formula for the result is <c>log(sum_(item,array) p(item,array) factor(item,array,index1,index2))</c>.</para>
/// </remarks>
/// <typeparam name="Distribution">The type of the distribution over an item.</typeparam>
public static double LogAverageFactor <Distribution>(T item, IArray2D <Distribution> array, int index1, int index2)
where Distribution : CanGetLogProb <T>
{
return(array[index1, index2].GetLogProb(item));
}
public Manhattan2DNode(short x, short y, IArray2D tab)
{
X = x;
Y = y;
Map = tab;
}
