function [class] = classifier(test, trainingSet, classifier, ...
    classifierInputs, simMetric, class0, inputs0, class1, ...
    inputs1)
% Classifies an object as being in class1 or class0
% Inputs:
% 1. test: object that needs to be classified.
% 2. trainingSet: matrix(n*m+1) of n objects, each of dimension m
%    The last column indicates if the object is in class 1 (1) or 
%    class 0 (0).
% 3. classifier: type of classifier to be used. Options so far:
%    a. 'kNN-simple': picks the class where the max number of k 
%       nearest neighbours are from.
%    b. 'mostRepElement': picks the class whose most representative
%       element is closest to the object.
%    c. (email)
% 4. classifierInputs: inputs to the classifer
%    a. kNN-simple: k
%    b. mostRepElement: ?
% 5. simMetric: type of similarity metric to use
% 6. class0: Name of the distribution followed by class 0
% 7. inputs0: parameters of class 0's distribution
% 8. class1: Name of the distribution followed by class 1
% 9. inputs1: parameters of class 1's distribution
m = length(test);
switch(classifier)
    case 'kNN-simple':
        numCols = size(trainingSet, 2);
        % Q. k-NN currently assumes that the point and the training set are
        % from the same class.
        [countNN linNN indexCount indexLin] = kNN(test, trainingSet(:,1:numCols-1), classifierInputs(1), class0, inputs0);
        indices = 0;
        if(simMetric == 'count')
            indices = indexCount;
        else 
            indices = indexLin;
        end
        count1 =0;
        count0 = 0;
        for i=1:size(indices,1)
            if(trainingSet(i,m) == 0)
                count0 = count0 + 1;
            else
                count1 = count1 + 1;
            end
        end
        if(count0 > count1)
            class = 0;
        else class = 1;
        end
    case 'mostRepElement'
        simMatrix = zeros(size(trainingSet,1), size(trainingSet,1));
        for i =1:size(trainingSet,1)
            for j=1:size(trainingSet,1)
                a = trainingSet(i,1:m);
                b= trainingSet(j,1:m);
                [AintersectB AnotB BnotA] = featurescomp(a,b);                
                if(simMetric == 'count')
                    simMatrix(i,j) = simCounting(AintersectB, AnotB, BnotA,  1 0.5 0.5);                                        
                elseif(simMetric == 'lin')
                    [probAintersectB] = calcProbIntersect(AintersectB
                end