import numpy as np
import k_medias as km
reader = np.genfromtxt("files/irisbin.csv", delimiter=',')
data = reader[:, 0:4]
y = reader[:, 4:]
H, W = data.shape
indexs = np.arange(0, H, 1)
np.random.shuffle(indexs)
k = 13
centroides, distancia, grupoCentroides = km.k_medias_tol(data,indexs, k, 0.05, 500)
varianzas = np.array([np.var(data[indexs[grupoCentroides == i], : ], axis = 0) for i in range(k)])
var = np.mean( np.mean(varianzas, axis = 1))
transfGauss = km.gauss_k_medias(data, centroides, k)
outputGauss = np.append(transfGauss, y, axis = 1)
np.savetxt("files/iris_k_medias_" + str(k) + ".csv", outputGauss, delimiter = ",")
reader = np.genfromtxt("files/merval.csv", delimiter=',')
#armar al base de datos
data = np.array([ reader[i * 6 : (i+1) * 6] for i in range(int(reader.shape[0] / 6)) ])
inData = data[:, 0:5]
yd = np.expand_dims(data[:,5], axis = 1)
H = inData.shape[0]
indexs = np.arange(0, H, 1)
np.random.shuffle(indexs)
minK = 15
maxK = 35
distPromV = np.ones((maxK-minK,1))
results = [km.k_medias_tol(inData, indexs, k, 0.05, 400, 200) for k in range(minK, maxK)]
for k in range(minK, maxK):
distPromV[k-minK] = results[k-minK][1]
fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(np.arange(minK, maxK), distPromV)
ax.scatter(np.arange(minK, maxK), distPromV)
ax.set_title("Curva k_medias Merval")
ax.set_xticks(np.arange(minK, maxK))
ax.set_xlabel(" K ")
import k_medias as km
from matplotlib import pyplot as plt
import multiprocessing as mp
np.random.seed(124394140)
reader = np.genfromtxt("files/XOR_trn.csv", delimiter=',')
data = reader[:, 0:2]
y = np.expand_dims(reader[:, 2], axis = 1)
H, W = data.shape
indexs = np.arange(0, H, 1)
np.random.shuffle(indexs)
k = 4
centroides, distProm, grupos = km.k_medias_tol(data,indexs, k)
for i in range(k):
plt.scatter(data[indexs[grupos == i], 0], data[indexs[grupos == i], 1])
plt.scatter(centroides[i][0], centroides[i][1], marker="^")
plt.show()
transfGauss = km.gauss_k_medias(data, centroides, k)
outputGauss = np.append(transfGauss, y, axis = 1)
#np.savetxt("files/xor_k_medias_" + str(k) + ".csv", outputGauss, delimiter = ",")