import project3 as p3
import random as ra
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as pat
from numpy import linalg as LA
#-------------------------------------------------------------------------------
# Section 1.a
#-------------------------------------------------------------------------------
# Test the provided kMeans method on the toy data for K=[1,2,3,4] with
# several different random initializations. Provide plots of the solution
# for each K that minimizes the total distortion cost.
X = p3.readData('toy_data.txt')
#for K in [1,2,3,4]:
# (Mu,P,Var) = p3.init(X,K)
# (Mu,P,Var,post) = p3.kMeans(X, K, Mu, P, Var)
#p3.plot2D(X,K,Mu,P,Var,post, "Title")
#-------------------------------------------------------------------------------
#-------------------------------------------------------------------------------
# Section 1.b
#-------------------------------------------------------------------------------
# Fill in the Estep, Mstep, and mixGauss methods.
# Test your Estep using K=3, after initializing using
# (Mu,P,Var) = p3.init(X,K,fixedmeans=True). The returned log-likelihood
# should be -1331.67489.
import project3 as p3
import random as ra
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as pat
from numpy import linalg as LA
#-------------------------------------------------------------------------------
# Section 1.a
#-------------------------------------------------------------------------------
# Test the provided kMeans method on the toy data for K=[1,2,3,4] with
# several different random initializations. Provide plots of the solution
# for each K that minimizes the total distortion cost.
X = p3.readData('toy_data.txt')
# for K in [1,2,3,4]:
# print "Calculating for K=", K
# (Mu, P, Var) = p3.init(X,K)
# (Mu,P,Var,post) = p3.kMeans(X, K, Mu, P, Var)
# p3.plot2D(X,K,Mu,P,Var,post,"plot for K = %d" % K)
#-------------------------------------------------------------------------------
#-------------------------------------------------------------------------------
# Section 1.b
#-------------------------------------------------------------------------------
# Fill in the Estep, Mstep, and mixGauss methods.
# Test your Estep using K=3, after initializing using
# (Mu,P,Var) = p3.init(X,K,fixedmeans=True). The returned log-likelihood
# should be -1331.67489.
############### Imports #############################
# TODO: insert your project 3 file name here:
import project3 as p3
import numpy as np
############### Read data ############################
X = p3.readData('matrixCompletionTest_incomplete.txt')
Xc = p3.readData('matrixCompletionTest_complete.txt')
K = 4
n, d = np.shape(X)
############### Functions #############################
def initToyFixed(X, K):
n, d = np.shape(X)
P = np.ones((K, 1)) / float(K)
Mu = np.array([[2.,3.,4.,0.,0.],\
[0.,0.,2.,3.,3.],\
[2.,5.,2.,3.,4.],\
[0.,5.,3.,4.,2.]])
Var = np.mean((X - np.tile(np.mean(X, axis=0), (n, 1)))**2) * np.ones(
(K, 1))
return (Mu, P, Var)
############### Init ##################################
Mu, P, Var = initToyFixed(X, K)
import project3 as p3
import random as ra
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as pat
from numpy import linalg as LA
X = p3.readData("toy_data.txt")
#-------------------------------------------------------------------------------
# Section 1.a
#-------------------------------------------------------------------------------
# Test the provided kMeans method on the toy data for K=[1,2,3,4] with
# several different random initializations. Provide plots of the solution
# for each K that minimizes the total distortion cost.
print "----------------------SECTION 1.A----------------------"
K_arr = [1, 2, 3, 4]
tries = 4
print X
for K in K_arr:
for t in range(tries):
(Mu, P, Var) = p3.init(X, K)
(Mu, P, Var, post) = p3.kMeans(X, K, Mu, P, Var)
title = 'Plot for K = ' + str(K) + ', Try ' + str(t)
p3.plot2D(X, K, Mu, P, Var, post, title)
#-------------------------------------------------------------------------------
#-------------------------------------------------------------------------------
# Section 1.b