/// <summary>
/// Creates a new <see cref="KNearestNeighbors"/> algorithm from an existing
/// <see cref="KDTree{T}"/>. The tree must have been created using the input
/// points and the point's class labels as the associated node information.
/// </summary>
///
/// <param name="tree">The <see cref="KDTree{T}"/> containing the input points and their integer labels.</param>
/// <param name="k">The number of nearest neighbors to be used in the decision.</param>
/// <param name="classes">The number of classes in the classification problem.</param>
/// <param name="inputs">The input data points.</param>
/// <param name="outputs">The associated labels for the input points.</param>
///
/// <returns>A <see cref="KNearestNeighbors"/> algorithm initialized from the tree.</returns>
///
public static KNearestNeighbors FromTree(KDTree <int> tree, int k, int classes, double[][] inputs, int[] outputs)
{
var knn = new KNearestNeighbors();
knn.K = k;
knn.Inputs = inputs;
knn.Outputs = outputs;
knn.NumberOfInputs = inputs.Columns();
knn.NumberOfOutputs = outputs.DistinctCount();
knn.NumberOfClasses = knn.NumberOfOutputs;
knn.tree = tree;
return(knn);
}
//Predict match for paraticipate of each user
public Hashtable Predict(int userID, SportEvent sevent)
{
Hashtable answersTable = new Hashtable();
//Building vector from data
double[][] test = new double[][]
{
new double[] { sevent.CategoryID, sevent.PlaceID }
};
foreach (DictionaryEntry s in usersTraining)
{
//foreach user except the creator of the user
if (!s.Key.Equals(userID))
{
//Checking for a match
Accord.MachineLearning.KNearestNeighbors model = s.Value as Accord.MachineLearning.KNearestNeighbors;
int[] answers = model.Decide(test);
bool answerBool = answers[0] != 0;
answersTable.Add(s.Key, answerBool);
Debug.WriteLine("User" + s.Key + ". This event is suitable for him: " + answers[0]);
}
}
return(answersTable);
}
/// <summary>
/// Creates a new <see cref="KNearestNeighbors"/> algorithm from an existing
/// <see cref="KDTree{T}"/>. The tree must have been created using the input
/// points and the point's class labels as the associated node information.
/// </summary>
///
/// <param name="tree">The <see cref="KDTree{T}"/> containing the input points and their integer labels.</param>
/// <param name="k">The number of nearest neighbors to be used in the decision.</param>
/// <param name="classes">The number of classes in the classification problem.</param>
/// <param name="inputs">The input data points.</param>
/// <param name="outputs">The associated labels for the input points.</param>
///
/// <returns>A <see cref="KNearestNeighbors"/> algorithm initialized from the tree.</returns>
///
public static KNearestNeighbors FromTree(KDTree <int> tree, int k, int classes, double[][] inputs, int[] outputs)
{
var knn = new KNearestNeighbors(tree, k, classes, inputs, outputs, tree.Distance);
return(knn);
}
