function [countNN indexCount] = kNN(test, dataBase, k)
% Determining the k nearest neighbors of a given point from 
% a given database using three similarity metrics- counting, lin and re.
% Assumption: 1. all points are from the same class
%             2. k < number of points in database
% Inputs:
% 1. point: vector in the same space as the objects in the database
% 2. dataBase: matrix (n*m) n= number of objects in the database; m =
% dimension of each object (note: point should have same dimension as all
% points in the database)
% 3. k: number of nearest neighbours
% 4. class: distribution that the objects in the database were created from
% 5. inputsDistrib: inputs to the above distribution.

% Outputs:
% 1. countNN: matrix (k*m)- k vectors of dimension m rep the k nearest
% neighbours, using counting similarity metric.
% 2. linNN: matrix (k*m)- k vectors of dimension m representing the k
% nearest neighbours, using lin similarity metric.
% 3. indexCount: indices of the kNN using simCounting
% 4. indexLin: indices of the kNN using simLin
simCount = zeros(size(dataBase,1),2);
simlin = zeros(size(dataBase,1),2);
simRe = zeros(1, length(dataBase));
for i= 1:size(dataBase,1)
    [AintersectB AnotB BnotA] = featurescomp(test, dataBase(i,:));
    %[probAnotB] = calcProbExclusive('AnotB', AnotB,class, class,...   
     %   inputsDistrib, inputsDistrib, 0);
    %[probBnotA] = calcProbExclusive('BnotA', BnotA,class, class,...
     %   inputsDistrib, inputsDistrib, 0);
    %[probAintersectB] = calcProbIntersect(AintersectB, class, class, inputsDistrib, inputsDistrib, 0);
    
    % hardcoding inputs to similarity metrics for now
    [simCount(i,:)] = [i simCounting(AintersectB, AnotB, BnotA, 1, 0.5, 0.5)];
    %[simlin(i,:)] = [i simLin(probAintersectB, probAnotB, probBnotA)];
end
simCount = sortrows(simCount,2);
%simCount
%simlin = sortrows(simlin,2);
indexCount = simCount(size(simCount, 1):-1:size(simCount,1)-k+1,1);
%indexLin = simlin(size(simlin, 1):-1:size(simlin,1)-k+1,1);
indexLin = indexCount;
%disp(simCount);
%disp(simlin);
j = 1;
for i=size(dataBase,1):-1:size(dataBase,1)-k+1
    countNN(j,:) = dataBase(simCount(i,1),:);
%    linNN(j,:) = dataBase(simlin(i,1),:);
    j= j+1;
end
linNN = countNN;