perceptron.m

%Perceptron training error rate is 0%.
%Perceptron test error rate is 20%.

%This function takes in a training data matrix Xtrain, training
%label vector ytrain and uses them to the decision rule corresponding
%to a very simple one-layer neural network: the perceptron. 
%It also takes in a data matrix Xrun and produces a vector of label
%guesses yguess, corresponding to the sign of the linear prediction.
function yguess = perceptron(Xtrain,ytrain,Xrun)

[row, col] = size(Xrun);
yguess = zeros(row, 1);

invX = pinv(Xtrain' * Xtrain);

for i = 1:row
    if (Xrun(i,:) * invX * Xtrain' * ytrain >= 0)
        yguess(i) = 1;
    else
        yguess(i) = -1;
    end
end

%Your code should go above this line.
if (~iscolumn(yguess))
    error("yguess is not a column vector.")
elseif (length(yguess)~=size(Xrun,1))
    error("Length of yguess is not equal to the number of rows in Xrun.")
elseif (sum(unique(abs(ytrain))~=1))
    warning("Some elements in ytrain are not +1 or -1.")
elseif (sum(unique(abs(yguess))~=1))
    warning("Some elements in yguess are not +1 or -1.")
end