def scan_points(points):
scans = {}
for pos in points:
X = points[pos]
(eps, min_pts) = get_params(pos)
scan = DBScan(eps, min_pts)
scan.fit(X)
scans[pos] = scan
return scans
def dbscan_estimate(self, pf):
eps = 0.5
minposes = 4
# Run DBSCAN
db = DBScan(pf.particlecloud, eps, minposes)
db.run()
# Get the estimate
guess = db.getguess()
self.largestClusterSize = db.largestClusterSize
return guess
def main():
data_iris = genfromtxt('iris-data.csv', delimiter=',')
data_iris_values = np.delete(data_iris, np.s_[4], 1)
data_iris = genfromtxt('iris-data.csv', delimiter=',', dtype=int)
data_iris_labels = np.delete(data_iris, np.s_[0:4], 1)
data_iris_labels = np.concatenate(data_iris_labels)
labels = DBScan(data_iris, 1, 15)
print(labels)
score = 0
maxscore = 0
for i in range(0, len(data_iris_labels)):
if (data_iris_labels[i] == labels[i]):
score += 1
maxscore += 1
accuracy = score / maxscore
print('Accuracy score (DBScan): %.2f' % accuracy)
plt.figure()
sum_squared_errors = []
for n_clusters in range(2, 10):
kmeans = KMeans(n_clusters=n_clusters, max_iter=100)
kmeans.fit(iris_data, normalize=True)
sse = kmeans.sum_squared_error()
sum_squared_errors.append(sse)
plt.plot(sum_squared_errors)
plt.xlabel('# of clusters')
plt.ylabel('SSE')
plt.show()
# TODO 7: DBSCAN nad Iris podacima, prikazati rezultate na grafiku isto kao kod K-means
dbscan = DBScan(epsilon=0.5, min_points=3)
dbscan.fit(iris_data)
colors = {0: 'red', 1: 'green', 2: 'blue'}
plt.figure()
for idx, cluster in enumerate(dbscan.clusters):
for datum in cluster.data: # iscrtavanje tacaka
plt.scatter(datum[0], datum[1], c=colors[idx])
plt.xlabel('Sepal width')
plt.ylabel('Petal length')
plt.show()
def test_dbscan(db, radius=0.3, min_pts=50):
dbscan = DBScan(db, radius, min_pts)
clusters = dbscan.cluster()
plot_clusters(clusters)
print('Found %d clusters' % len(clusters))
data.append((x1, y1))
data.append((x2, y2))
plt.show()
# TODO 5: K-means nad ovim podacima
kmeans = KMeans(n_clusters=2, max_iter=100)
kmeans.fit(data)
colors = {0: 'red', 1: 'green'}
plt.figure()
for idx, cluster in enumerate(kmeans.clusters):
plt.scatter(cluster.center[0], cluster.center[1], c=colors[idx], marker='x', s=200) # iscrtavanje centara
for datum in cluster.data: # iscrtavanje tacaka
plt.scatter(datum[0], datum[1], c=colors[idx])
plt.show()
# TODO 7: DBSCAN nad ovim podacima
dbscan = DBScan(epsilon=1.2, min_points=3)
dbscan.fit(data)
colors = {0: 'red', 1: 'pink', 2: 'yellow', 3: 'cyan', 4: 'green', 5: 'blue'}
plt.figure()
for idx, cluster in enumerate(dbscan.clusters):
for datum in cluster.data: # iscrtavanje tacaka
plt.scatter(datum[0], datum[1], c=colors[idx%6])
plt.show()
class Main:
exp_execute = 4
############################################
# Experiment DBScan 1
# Verify distances with and without scale
############################################
if exp_execute == 1:
for d, scale in zip([0, 0, 1, 1], [False, True, False, True]):
datasets = read_datasets()
if scale:
datasets = scale_datasets(datasets)
datasets = split_data(datasets)
dist = 1.0
min_neighb = 3
model = DBScan(distance=dist, min_neighbors=min_neighb)
dataset = datasets[d]
res = model.fit(dataset['train'].to_numpy())
res = model.get_description_dist()
print(res)
############################################
# Experiment DBScan 2
# Execute model for Dataset 1
############################################
if exp_execute == 2:
datasets = read_datasets()
datasets = scale_datasets(datasets)
datasets = split_data(datasets)
for dist in [0.05, 0.1, 0.15]:
df_merge = datasets[0]['train'].copy()
df_merge['Type'] = -2
df_merge['Cluster'] = -2
df_merge['Experiment'] = 'remove'
for min_neighb in [3, 5, 8]:
model = DBScan(distance=dist, min_neighbors=min_neighb)
dataset = datasets[0]
res = model.fit(dataset['train'].to_numpy())
pred = model.predict(dataset['train'].to_numpy(), res,
dataset['test'].to_numpy())
# plot(dataset['train'].copy(), res, file_name=f'dbscan_vardist_{dist}_{min_neighb}.png')
write_results(datasets,
f'dbscan_vardist_{dist}_{min_neighb}_datasets')
write_results(res, f'dbscan_vardist_{dist}_{min_neighb}_res')
write_results(pred, f'dbscan_vardist_{dist}_{min_neighb}_pred')
df_merge = pd.concat([
df_merge,
merge_result(dataset['train'].copy(), res,
f'Distance: {dist} Neighbor: {min_neighb}')
])
plot_all(df_merge, file_name=f'all_dbscan_{dist}.png')
############################################
# Experiment DBScan 3
# Execute model for Dataset 1
############################################
if exp_execute == 3:
datasets = read_datasets()
datasets = scale_datasets(datasets)
datasets = split_data(datasets)
dist = 0.1
min_neighb = 3
model = DBScan(distance=dist, min_neighbors=min_neighb)
dataset = datasets[0]
res = model.fit(dataset['train'].to_numpy())
pred = model.predict(dataset['train'].to_numpy(), res,
dataset['test'].to_numpy())
plot_pred(dataset['test'].copy(),
pred,
file_name=f'dbscan_pred_{dist}_{min_neighb}.png')
############################################
# Experiment DBScan 4
# Execute model for Dataset 2
############################################
if exp_execute == 4:
datasets = read_datasets()
datasets = scale_datasets(datasets)
datasets = split_data(datasets)
for dist in [0.05, 0.1, 0.15]:
df_merge = datasets[1]['train'].copy()
df_merge['Type'] = -2
df_merge['Cluster'] = -2
df_merge['Experiment'] = 'remove'
for min_neighb in [3, 5, 8]:
model = DBScan(distance=dist, min_neighbors=min_neighb)
dataset = datasets[1]
res = model.fit(dataset['train'].to_numpy())
pred = model.predict(dataset['train'].to_numpy(), res,
dataset['test'].to_numpy())
# plot(dataset['train'].copy(), res, file_name=f'dbscan_vardist_{dist}_{min_neighb}.png')
pca = PCA(n_components=2, random_state=42)
df_pca = pca.fit_transform(['train'].copy())
write_results(
datasets, f'dbscan_vardist_{dist}_{min_neighb}_datasets_2')
write_results(res, f'dbscan_vardist_{dist}_{min_neighb}_res_2')
write_results(pred,
f'dbscan_vardist_{dist}_{min_neighb}_pred_2')
df_merge = pd.concat([
df_merge,
merge_result(df_pca, res,
f'Distance: {dist} Neighbor: {min_neighb}')
])
plot_all(df_merge, file_name=f'all_dbscan_{dist}_2.png')