meanrow = X.mean(0)

# present the row vector as an image

plt.figure(figsize=(3,3))

plt.imshow(np.reshape(meanrow,(28,28)), cmap=plt.cm.binary)

plt.title('Mean image of ' + log)

"""

Read input-vector (image) and target class (label, 0-9) and return it as list of tuples.

Adapted from: https://martin-thoma.com/classify-mnist-with-pybrain/

"""

# Open the images with gzip in read binary mode

images = open(imagefile, 'rb')

labels = open(labelfile, 'rb')

# Read the binary data

# We have to get big endian unsigned int. So we need '>I'

# Get metadata for images

images.read(4) # skip the magic_number

number_of_images = images.read(4)

number_of_images = unpack('>I', number_of_images)[0]

rows = images.read(4)

rows = unpack('>I', rows)[0]

cols = images.read(4)

cols = unpack('>I', cols)[0]

# Get metadata for labels

labels.read(4) # skip the magic_number

N = labels.read(4)

N = unpack('>I', N)[0]

if number_of_images != N:

raise Exception('number of labels did not match the number of images')

# Get the data

X = np.zeros((N, rows * cols), dtype=np.uint8) # Initialize numpy array

y = np.zeros(N, dtype=np.uint8) # Initialize numpy array

for i in range(N):

for id in range(rows * cols):

tmp_pixel = images.read(1) # Just a single byte

tmp_pixel = unpack('>B', tmp_pixel)[0]

X[i][id] = tmp_pixel

tmp_label = labels.read(1)

y[i] = unpack('>B', tmp_label)[0]

# =======================================

# Complete the code here.

# return scores like this: return [score, score, score, score]

# =======================================

from sklearn.cluster import KMeans

#print('fuck X ', X.shape)

#print('fuck y ', y.shape)

clf = KMeans(n_clusters)

clf.fit(X)

from sklearn import metrics

ari = metrics.adjusted_rand_score(y, clf.labels_)

mri = metrics.adjusted_mutual_info_score(y, clf.labels_)

v_measure = metrics.v_measure_score(y, clf.labels_)

'''

silhouette_coeff = metrics.silhouette_score(X, clf.labels_,

metric='euclidean',

sample_size=300)

'''

silhouette_coeff = metrics.silhouette_score(X, clf.labels_)

show_images(n_clusters, clf, pca)

arr = []

sum = 0

for i in range(0, len(eigenvalues)):

sum += eigenvalues[i]

cumulate = 0

for i in range(0, len(eigenvalues)):

cumulate += eigenvalues[i]

arr.append(cumulate / sum)

# in kmean.cluster_centers_.shape, what we get is a (n_clusters, 84) array

# so for every [1,84] vector we have inside kmean.cluster_centers_,

# we need to trasform it back to the original vector

# then transorm it back into the original 28x28 matrix

cluster_centers_org = pca.inverse_transform(kmean.cluster_centers_)

n_row = 1

n_col = n_clusters

plt.figure(figsize=(1.8 * n_col, 2.4 * n_row))

plt.subplots_adjust(bottom=0, left=.01, right=.99, top=.90, hspace=.35)

for i in range(n_row * n_col):

plt.subplot(n_row, n_col, i + 1)

plt.imshow(cluster_centers_org[i].reshape(28,28), cmap=plt.cm.gray)

title_text = 'Center ' + str(i + 1)

plt.title(title_text, size=12)

plt.xticks(())

plt.yticks(())

plt.savefig('ex2_n_clusters_' + str(n_clusters) + '.png')

# Load the dataset

fname_img = 't10k-images.idx3-ubyte'

fname_lbl = 't10k-labels.idx1-ubyte'

[X, y]=get_labeled_data(fname_img, fname_lbl)

# Dataset info

print('Total dataset size: ')

# each row of matrix X represents an image

print("n_samples: ", X.shape[0])

print("n_features: ", X.shape[1])

# Plot the mean image

plot_mean_image(X, 'all images')

# =======================================

# Complete the code here.

# Use PCA to reduce the dimension here.

# You may want to use the following codes. Feel free to write in your own way.

# - pca = PCA(n_components=...)

# - pca.fit(X)

# - print('We need', pca.n_components_, 'dimensions to preserve 0.95 POV')

# =======================================

pca = PCA(n_components=0.95).fit(X)

X_pca = pca.transform(X)

# Clustering

range_n_clusters = [8, 9, 10, 11, 12]

ari_score = [None] * len(range_n_clusters)

mri_score = [None] * len(range_n_clusters)

v_measure_score = [None] * len(range_n_clusters)

silhouette_avg = [None] * len(range_n_clusters)

print(79 * '_')

print('% 9s' % 'n_clusters'

' time ARI MRI v-measure score avg silhouette score')

for n_clusters in range_n_clusters:

t0 = time()

i = n_clusters - range_n_clusters[0]

#print("Number of clusters is: ", n_clusters)

[ari_score[i], mri_score[i], v_measure_score[i], silhouette_avg[i]] = my_clustering(X_pca, y, n_clusters, pca)

# print('The ARI score is: ', ari_score[i])

# print('The MRI score is: ', mri_score[i])

# print('The v-measure score is: ', v_measure_score[i])

# print('The average silhouette score is: ', silhouette_avg[i])

print('% 9s %.15f %.15f %.15f %.15f %.15f'

% (n_clusters, (time() - t0), ari_score[i], mri_score[i], v_measure_score[i], silhouette_avg[i]))

print(79 * '_')

# =======================================

# Complete the code here.

# Plot scores of all four evaluation metrics as functions of n_clusters.

# =======================================

# Plot Adjusted Rand Index

plt.figure()

plt.plot(range_n_clusters, ari_score, color='b', linestyle='-',label='ARI')

plt.title('Adjusted Rand Index')

plt.ylabel('score')

plt.xlabel('number of clusters')

#plt.legend(loc='upper left', prop={'size':6})

plt.savefig('ex2_score_ARI.png')

# Plot Mutual Information based scores

plt.figure()

plt.plot(range_n_clusters, mri_score, color='r', linestyle='-',label='MRI')

plt.title('Mutual Information based scores')

plt.ylabel('score')

plt.xlabel('number of clusters')

#plt.legend(loc='upper left', prop={'size':6})

plt.savefig('ex2_score_MRI.png')

# Plot V-measure

plt.figure()

plt.plot(range_n_clusters, v_measure_score, color='c', linestyle='-',label='V-measure')

plt.title('V-measure')

plt.ylabel('score')

plt.xlabel('number of clusters')

#plt.legend(loc='upper left', prop={'size':6})

plt.savefig('ex2_score_V_measure.png')

# Plot Silhouette Coefficient

plt.figure()

plt.plot(range_n_clusters, silhouette_avg, color='m', linestyle='-',label='Silhouette')

plt.title('Silhouette Coefficient')

plt.ylabel('score')

plt.xlabel('number of clusters')

#plt.legend(loc='upper left', prop={'size':6})

plt.savefig('ex2_score_Silhouette.png')

plt.figure()

plt.plot(range_n_clusters, ari_score, color='b', linestyle='-',label='ARI')

plt.plot(range_n_clusters, mri_score, color='r', linestyle='-',label='MRI')

plt.plot(range_n_clusters, v_measure_score, color='c', linestyle='-',label='V-measure')

plt.plot(range_n_clusters, silhouette_avg, color='m', linestyle='-',label='Silhouette')

plt.title('Overall Scores')

plt.ylabel('score')

plt.xlabel('number of clusters')

plt.legend(loc='upper left', prop={'size':6})

plt.savefig('ex2_score_overall.png')

#plt.show()

# v-measure: range(0,1) 1 means perfectly complete labeling

# ARI adjusted Rand index: is ensured to have a value close to 0.0

# for random labeling independently of the

# number of clusters and samples and exactly 1.0

# when the clusterings are identical (up to a

# permutation).

# Similarity score between -1.0 and 1.0.

# Random labelings have an ARI close to 0.0. 1.0 stands for perfect match.

# The AMI: returns a value of 1 when the two partitions are identical (ie perfectly matched).

# Random partitions (independent labellings) have an expected AMI around 0 on average hence can be negative.

# Silhoutte: The best value is 1 and the worst value is -1.

# Values near 0 indicate overlapping clusters.

# Negative values generally indicate that a sample has been assigned