Logistic prediction

log_regression.py

def sigmoid(X):
    '''Compute the sigmoid function '''
    #d = zeros(shape=(X.shape))

    den = 1.0 + e ** (-1.0 * X)

    d = 1.0 / den

    return d


def compute_cost(theta, X, y):
    '''
    Comput cost for logistic regression
    '''
    #Number of training samples

    theta.shape = (1, 3)

    m = y.size

    h = sigmoid(X.dot(theta.T))

    h[h==1] = 0.99999999999

    # according to wikipedia, the linear version is:
    # sigmoid^y  * (1-sigmoid)^(1-y)
    J = ((y.T.dot(log(h))) + ((1.0 - y.T).dot(log(1.0 - h))))

    # not really sure why negative, but in order to make it a COST function and not something to maximize, need it
    return -(1.0/m) * J.sum()


def compute_grad(theta, X, y):

    #print theta.shape

    theta.shape = (1, 3)

    m = y.size

    grad = zeros(3)

    h = np.squeeze(sigmoid(X.dot(theta.T)))

    delta = h - y

    l = grad.size

    for i in range(l):
        sumdelta = delta.T.dot(X[:, i])
        grad[i] = (1.0 / m) * sumdelta * - 1

    theta.shape = (3,)

    return  grad