# database = DatabaseIris()
# a = ['Compound.txt' , 'flame.txt' ,'D31.txt', 'jain.txt' , 'pathbased.txt' , 'R15.txt' , 'spiral.txt']
#a = ['Aggregation.txt' ]
# a = ['flame.txt', 'jain.txt']
#a = [('spiral.txt', 3)]
#a = [('R15.txt', 15)]
#a = [('Compound.txt', 6)]
#a = [('pathbased.txt', 3)]
a = [('flame.txt', 2)]
results = []
for f, K in a:
fi = '../datasets/{}'.format(f)
database = TwoDimensionData(fi, '\t')
base_name = os.path.basename(fi)
name = os.path.splitext(base_name)[0]
for rho in np.arange(1.0, 3.4, 0.2):
# for rho in [2.8]:
dissimilarity = DensityDistance(rho=rho)
cluster = Cluster(database, dissimilarity, P_size=50, K=K, max_iterations=50)
score, score_normalized = cluster.compute()
results.append((name, K, rho, score, score_normalized))
for name, K, rho, score, score_normalized in results:
import numpy as np
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
from sklearn import metrics
from Dissimilarity import DensityDistance
from Database import TwoDimensionData
database = TwoDimensionData("../datasets/D31.txt", '\t')
#dissimilarity = DensityDistance(rho=1.2)
data = np.asarray(database.data)
labels = np.asarray(database.labels)
kmeans = KMeans(n_clusters=31, random_state=1).fit_predict(data)
metrics.silhouette_score(data, labels, metric='euclidean')
plt.scatter(data[:, 0], data[:, 1], c=kmeans)
plt.title("K-means")
plt.show()