In [2]:

    

using Gadfly


    

In [3]:

    

data = readdlm("../../../datasets/ml-ex1/ex1data2.txt", ',');


    

In [4]:

    

oX = data[:,1:2];
y = data[:,3];
m = length(y);
X = [ones(m, 1) oX];


    

In [25]:

    

function feature_normalize(X)
    means = mean(X, 1)
    stds = std(X, 1)
    return (X .- means) ./ stds
end

function compute_cost(X, y, theta)
  return sum(((theta' * X')' - y).^2) / (2 * length(y))
end

function gradient_descent(X, y, theta, alpha, iterations)
  cost_history = zeros(iterations, 1)
  for i in collect(1:iterations)
    theta = (theta' - ((alpha / length(y)) * sum((sum((theta' .* X), 2) - y).*X, 1)))'
    cost_history[i] = compute_cost(X, y, theta)
  end
  return theta, cost_history
end


    

    
Out[25]:





gradient_descent (generic function with 1 method)

In [29]:

    

X = [ones(m, 1) feature_normalize(oX)]
num_features = length(X[1,:]);
theta = zeros(num_features, 1);
iterations = 50;
alpha = 0.1;
initial_cost = compute_cost(X, y, theta);
(ptheta, cost_history) = gradient_descent(X, y, theta, alpha, iterations);

plot(x=1:iterations, y=cost_history[:,1], Geom.line)


    

    
Out[29]:

In [33]:

    

function normal_equation(X, y)
    return (pinv(X'*X))*(X'*y)
end

normal_equation(X, y)


    

    
Out[33]:





3-element Array{Float64,1}:
 340413.0 
 110631.0 
  -6649.47

In [ ]: