def main():
if len(sys.argv) == 1:
print("Ejecuta: python3 example2.py [p].")
return
p = float(sys.argv[1])
X = utils.gen_points3()
(x, y) = (X[:, 0], X[:, 1])
sm_x = np.mean(x)
sm_y = np.mean(y)
m_d = np.array([sm_x, sm_y], dtype = 'float')
gm = optimization.gen_mean(X, m_d, p)
Y = X.copy()
no_eig = 1
Y[:, 0] -= sm_x
Y[:, 1] -= sm_y
W = optimization.PCA(X, no_eig)
_, V = optimization.PCAGM(X, gm, no_eig, W, p)
print("PCA (primera direccion):")
print(W)
print("PCAGM (primera direccion):")
print(V)
plt.plot(Y[:, 0], Y[:, 1], 'ro', color = 'g')
plt.quiver([0], [0], [W[0, 0]], [W[1, 0]], angles = 'xy', scale_units = 'xy', scale = 1, color = ['black'])
plt.quiver([0], [0], [V[0, 0]], [V[1, 0]], angles = 'xy', scale_units = 'xy', scale = 1, color = ['b'])
plt.axis([-2, 7, -2, 7])
plt.show()
def main():
if len(sys.argv) == 1:
print("Ejecuta: python3 example1.py [p].")
return
p = float(sys.argv[1])
p1 = np.random.multivariate_normal([0, 0], [[0.5, 0], [0, 0.5]], 100)
p2 = np.random.multivariate_normal([5, 5], [[0.3, 0], [0, 0.3]], 10)
data = np.concatenate((p1, p2), axis=0)
(x, y) = (p1[:, 0], p1[:, 1])
x = np.append(x, p2[:, 0])
y = np.append(y, p2[:, 1])
sm_x = np.mean(x)
sm_y = np.mean(y)
m_d = np.array([sm_x, sm_y], dtype='float')
gm = optimization.gen_mean(data, m_d, p)
print("Media muestral:")
print(m_d)
print("Media generalizada:")
print(gm)
plt.plot(x, y, 'ro')
plt.plot(sm_x, sm_y, 'bs', color='g')
plt.plot(gm[0], gm[1], 'bs')
plt.axis([-2, 7, -2, 7])
plt.show()
def main():
if len(sys.argv) == 1:
print("Ejecuta: python3 example3.py [p].")
return
p = float(sys.argv[1])
X = utils.gen_points_ex()
X_wo = utils.gen_points_no_out()
no_eig = 1
(x, y) = (X[:, 0], X[:, 1])
sm_x = np.mean(x)
sm_y = np.mean(y)
m_d = np.array([sm_x, sm_y], dtype='float')
gm = optimization.gen_mean(X, m_d, p)
W = optimization.PCA(X, no_eig)
W_wo = optimization.PCA(X_wo, no_eig)
W_0 = np.matrix('0.86607; 0.49991', dtype='float')
_, V = optimization.PCAGM(X, gm, 1, W_0, p)
PCA_fn = optimization.PCA_J_fn(X, W)
print("Resultados sin datos atipicos (PCA).")
print(W_wo)
print("Resultados con datos atipicos (PCA).")
print(W)
print("Resultados con datos atipicos (PCAGM).")
print(V)
plt.plot(X[:, 0], X[:, 1], 'ro', color='g')
plt.quiver([0], [0], [W[0, 0]], [W[1, 0]],
angles='xy',
scale_units='xy',
scale=1,
color=['black'])
plt.quiver([0], [0], [V[0, 0]], [V[1, 0]],
angles='xy',
scale_units='xy',
scale=1,
color=['b'])
plt.axis([-3, 5, -2, 7])
plt.show()
return 0
def main():
if len(sys.argv) == 1:
print("Ejecuta: python3 clustering.py [n] [p].")
return
n_proj = int(sys.argv[1])
p = float(sys.argv[2])
x_train, t_train, _, _ = mnist.load()
n, m = 360, 784
data = np.zeros((n, m), dtype='float')
labels = np.zeros((n, 1), dtype='int')
cnt, idx = 0, 0
while cnt < 300:
if t_train[idx] == 3 or t_train[idx] == 8 or t_train[idx] == 9:
data[cnt, :] = (x_train[idx, :] / LA.norm(x_train[idx, :], 1))
labels[cnt] = t_train[idx]
cnt += 1
idx += 1
while cnt < 360:
if t_train[idx] != 3 and t_train[idx] != 8 and t_train[idx] != 9:
data[cnt, :] = (x_train[idx, :] / LA.norm(x_train[idx, :], 1))
labels[cnt] = t_train[idx]
cnt += 1
idx += 1
W_PCA = optimization.PCA(data, n_proj)
data_proj_PCA = data.dot(W_PCA)
data_mean = np.zeros((data.shape[1]), dtype='float')
for i in range(data.shape[1]):
data_mean[i] = np.mean(data[:, i])
gm = optimization.gen_mean(data, data_mean, p)
_, W_PCAGM = optimization.PCAGM(data, gm, n_proj, W_PCA, p)
data_proj_PCAGM = data.dot(W_PCAGM)
k = 3
prec_1 = make_clustering(data_proj_PCA, labels, 3)
prec_2 = make_clustering(data_proj_PCAGM, labels, 3)
print("Precision de la clasificacion (PCA): ", prec_1)
print("Precision de la clasificacion (PCA GM): ", prec_2)
def main():
if len(sys.argv) == 1:
print("Ejecuta: python4 example1.py [p].")
return
p = float(sys.argv[1])
X = utils.gen_points_ex()
Y = X.copy()
(x, y) = (X[:, 0], X[:, 1])
sm_x = np.mean(x)
sm_y = np.mean(y)
m_d = np.array([sm_x, sm_y], dtype = 'float')
gm = optimization.gen_mean(X, m_d, p)
X[:, 0] -= gm[0]
X[:, 1] -= gm[1]
meanPoints = np.mean(Y, axis = 0)
Y = Y - meanPoints
linf, lsup = 0.0, 180.0
n = 150
dx = (lsup - linf) / float(n - 1)
angle = np.zeros((n), dtype = 'float')
y_PCA = np.zeros((n), dtype = 'float')
y_PCAGM = np.zeros((n), dtype = 'float')
for i in range(n):
angle[i] = float(linf) + i * dx
r_aux = np.radians(angle[i])
W_aux = np.zeros((2, 1), dtype = 'float')
W_aux[0, 0] = np.cos(r_aux)
W_aux[1, 0] = np.sin(r_aux)
y_PCA[i] = optimization.PCA_J_fn(Y, W_aux)
y_PCAGM[i] = optimization.PCAGM_J_fn(X, W_aux, p)
plt.plot(angle, y_PCA, color = 'g', label = 'Funcion de costo de PCA')
plt.plot(angle, y_PCAGM, color = 'b', label = 'Funcion de costo de PCA robusto')
plt.legend(bbox_to_anchor = (0., 1.02, 1., .102), loc = 3, ncol = 2, mode = "expand", borderaxespad = 0.)
plt.xlabel('Angulo')
plt.ylabel('Valor de la funcion de costo')
plt.show()