-
Notifications
You must be signed in to change notification settings - Fork 0
/
k-means.py
243 lines (221 loc) · 8.37 KB
/
k-means.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
import csv
import sys
import math
import random
import matplotlib.pyplot as plt
import numpy
class Cluster(object):
def __init__(self, centroid):
self.centroid = centroid
self.members = []
def addEntry(self, entry):
self.members.append(entry)
def changeCentroid(self, centroid):
self.centroid = centroid
def getWCScore(self, isManhattan):
score = 0
for entry in self.members:
dist = distance(self.centroid, entry, isManhattan)
dist = math.pow(dist, 2)
score += dist
return score
def loadCsv(filename): # we eliminate all unnecessary attributes
lines = csv.reader(open(filename, "rb"))
next(lines, None)
dataset = list(lines)
entryLength = len(dataset[0])
for i in range(len(dataset)):
attrNum = 0
while (attrNum < entryLength):
if (attrNum < 3):
dataset[i].pop(0)
elif (attrNum == 5):
dataset[i].pop(2)
elif (attrNum > 7):
dataset[i].pop(4)
attrNum += 1
for index, val in enumerate(dataset[i]):
dataset[i][index] = float(val)
return dataset
def distance(first, second, isManhattan):
result = 0
if isManhattan:
for index, val in enumerate(first):
result += val - second[index]
else:
sum = 0
for index, val in enumerate(first):
sum += math.pow(val-second[index], 2)
result = math.sqrt(sum)
return result
def generateRandomCentroids(dataset, k):
points = random.sample(dataset, k)
clusters = []
for centroid in points:
cluster = Cluster(centroid)
clusters.append(cluster)
return clusters
def connectToClusters(dataset, clusters, isManhattan):
for entry in dataset:
minDistance = 9999999.99
resultCentroidIndex = 0
for index, cluster in enumerate(clusters):
dist = distance(cluster.centroid, entry, isManhattan)
if ( dist < minDistance ):
minDistance = dist
resultCentroidIndex = index
clusters[resultCentroidIndex].addEntry(entry)
def changeClustersWithData(clusters):
for idx, cluster in enumerate(clusters):
newMeanCentroid = [0, 0, 0, 0]
for entry in cluster.members:
for index, val in enumerate(entry):
newMeanCentroid[index] += val/len(cluster.members)
clusters[idx].changeCentroid(newMeanCentroid)
def changeMembersWithNewClusters(clusters, isManhattan):
numOfChanges = 0
for mainIndex, mainCluster in enumerate(clusters):
for entry in mainCluster.members:
currentDistance = distance(mainCluster.centroid, entry, isManhattan)
minDistance = currentDistance
resultClusterIndex = mainIndex
for subIndex, subCluster in enumerate(clusters):
dist = distance(subCluster.centroid, entry, isManhattan)
if dist < minDistance:
minDistance = dist
resultClusterIndex = subIndex
if resultClusterIndex != mainIndex:
mainCluster.members.remove(entry)
clusters[resultClusterIndex].members.append(entry)
numOfChanges += 1
return numOfChanges
def loggedLoadCsv(filename): # we eliminate all unnecessary attributes
lines = csv.reader(open(filename, "rb"))
next(lines, None)
dataset = list(lines)
entryLength = len(dataset[0])
for i in range(len(dataset)):
attrNum = 0
while (attrNum < entryLength):
if (attrNum < 3):
dataset[i].pop(0)
elif (attrNum == 5):
dataset[i].pop(2)
elif (attrNum > 7):
dataset[i].pop(4)
attrNum += 1
for index, val in enumerate(dataset[i]):
if index == 2 | index == 3:
dataset[i][index] = math.log(float(val))
else:
dataset[i][index] = float(val)
return dataset
def standardizedLoadCsv(filename): # we eliminate all unnecessary attributes
lines = csv.reader(open(filename, "rb"))
next(lines, None)
dataset = list(lines)
entryLength = len(dataset[0])
for i in range(len(dataset)):
attrNum = 0
while (attrNum < entryLength):
if (attrNum < 3):
dataset[i].pop(0)
elif (attrNum == 5):
dataset[i].pop(2)
elif (attrNum > 7):
dataset[i].pop(4)
attrNum += 1
for index, val in enumerate(dataset[i]):
dataset[i][index] = float(val)
meanSet = []
stdSet = []
totalArray = {0:[],1:[],2:[],3:[]}
for entry in dataset:
for index,attrVal in enumerate(entry):
totalArray[index].append(attrVal)
for index,attr in enumerate(totalArray):
meanSet.append(numpy.mean(totalArray[index]))
stdSet.append(numpy.std(totalArray[index]))
for index, entry in enumerate(dataset):
for aIndex, attrVal in enumerate(dataset[index]):
dataset[index][aIndex] = abs( (dataset[index][aIndex]-meanSet[aIndex])/stdSet[aIndex] )
return dataset
def plotLat(clusters, isReview):
for oneCluster in clusters:
colorNum = numpy.random.rand(3,1)
if isReview:
plt.scatter(oneCluster.centroid[2], oneCluster.centroid[3], marker = "x", c=colorNum, s=100)
else:
plt.scatter(oneCluster.centroid[0], oneCluster.centroid[1], marker="x", c=colorNum, s=100)
xValues = []
yValues = []
for entry in oneCluster.members:
if isReview:
xValues.append(entry[2])
yValues.append(entry[3])
else:
xValues.append(entry[0])
yValues.append(entry[1])
plt.scatter(xValues, yValues, marker="o", c=colorNum, s=25)
if isReview:
plt.xlabel("Review Count")
plt.ylabel("Checkins")
else:
plt.xlabel("Lattitude")
plt.ylabel("Logntitude")
plt.show()
def main():
kNumber = int(sys.argv[2])
clusteringOption = int(sys.argv[3])
plotOption = sys.argv[4]
if clusteringOption == 2:
dataset = loggedLoadCsv(sys.argv[1])
elif clusteringOption == 3:
dataset = standardizedLoadCsv(sys.argv[1])
else:
dataset = loadCsv(sys.argv[1]) #argv1
clusters = generateRandomCentroids(dataset, kNumber)
if clusteringOption == 1:
isManhattan = False
connectToClusters(dataset, clusters, isManhattan)
changeClustersWithData(clusters)
while changeMembersWithNewClusters(clusters, isManhattan) != 0:
changeClustersWithData(clusters)
totalScore = 0
for index, cluster in enumerate(clusters):
totalScore += cluster.getWCScore(isManhattan)
print "WC-SSE=" + str(totalScore)
for index, cluster in enumerate(clusters):
print "Centroid" + str(index + 1) + "=" + str(cluster.centroid)
elif clusteringOption == 4:
isManhattan = True
connectToClusters(dataset, clusters, isManhattan)
changeClustersWithData(clusters)
while changeMembersWithNewClusters(clusters, isManhattan) != 0:
changeClustersWithData(clusters)
totalScore = 0
for index, cluster in enumerate(clusters):
totalScore += cluster.getWCScore(isManhattan)
print "WC-SSE=" + str(totalScore)
for index, cluster in enumerate(clusters):
print "Centroid" + str(index + 1) + "=" + str(cluster.centroid)
elif clusteringOption == 5:
percentNum = int(len(dataset) * 0.01)
downSampled = random.sample(dataset, percentNum)
clusters = generateRandomCentroids(downSampled, kNumber)
isManhattan = False
connectToClusters(downSampled, clusters, isManhattan)
changeClustersWithData(clusters)
while changeMembersWithNewClusters(clusters, isManhattan) != 0:
changeClustersWithData(clusters)
totalScore = 0
for index, cluster in enumerate(clusters):
totalScore += cluster.getWCScore(isManhattan)
print "WC-SSE=" + str(totalScore)
for index, cluster in enumerate(clusters):
print "Centroid" + str(index + 1) + "=" + str(cluster.centroid)
if plotOption == "1":
plotLat(clusters, False)
if plotOption == "2":
plotLat(clusters, True)
main()