matrix_factorization_compare.py

def matrix_factorization_quux(
    D, P, Q, steps=5000, alpha=0.0002, beta=0.02):
  K = P.shape[1]
  P = np.copy(P)
  Q = np.copy(Q)
  for step in xrange(steps):
    for i in xrange(len(D)):
      for j in xrange(len(D[i])):
        if not getmask(D)[i, j]:
          eij = D[i, j] - np.dot(P[i, :], Q[:, j])
          for k in xrange(K):
            P[i, k] = P[i, k] + alpha * (2 * eij * Q[k, j] - beta * P[i, k])
            Q[k, j] = Q[k, j] + alpha * (2 * eij * P[i, k] - beta * Q[k, j])
  return P, Q


if __name__ == '__main__':
  D = np.array([[5,  3, -1, 1],
                [4, -1, -1, 1],
                [1,  1, -1, 5],
                [1, -1, -1, 4],
                [-1, 1,  5, 5]])
  D = ma.masked_array(D, mask=D==-1)
  m, n = D.shape
  K = 2
  P = np.random.rand(m, K)
  Q = np.random.rand(K, n)

  np.set_printoptions(formatter={'all': lambda x: str(x).rjust(2)})
  print 'Ratings Matrix\n', D, '\n'

  np.set_printoptions(precision = 2, formatter=None)

  P_theano_bgd, Q_theano_bgd = matrix_factorization_bgd(D, P, Q)
  print 'Theano Batch Gradient Descent\n',\
    np.dot(P_theano_bgd, Q_theano_bgd), '\n'

  P_theano_sgd, Q_theano_sgd = matrix_factorization_sgd(D, P, Q)
  print 'Theano Stochastic Gradient Descent\n',\
    np.dot(P_theano_sgd, Q_theano_sgd), '\n'

  P_quux, Q_quux = matrix_factorization_quux(D, P, Q)
  print 'quuxlabs\n', np.dot(P_quux, Q_quux), '\n'