def draw_graph(datasets, n_clusters=2):
plot_num = 1
alg_names = []
eps_list = []
for eps in np.arange(0.1, 1, step=0.1):
alg_names.append('DBScan eps=%.1f' % eps)
eps_list.append(eps)
a = np.arange(len(alg_names)*2)+1
indices = np.hstack(a.reshape(len(alg_names), 2).T)
elice_utils.draw_init()
for dataset in datasets:
y = dataset['0']
X = np.array(dataset[['1', '2']])
for alg_name, eps in zip(alg_names, eps_list):
dbscan_result = run_DBScan(X, eps)
elice_utils.draw_graph(X, dbscan_result, alg_name, plot_num, len(alg_names), indices)
plot_num += 1
print(elice_utils.show_graph())
return dbscan_result
def draw_graph(datasets, n_clusters=2):
# DB-scan 알고리즘에서 eps값이 변화함에 따라 같은 클러스터로 인식하는 범위가 달라지는 여부를 비교해볼 수 있다.
# 예상대로 eps값이 커질수록 멀리 떨어진 부분을 같은 클러스터로 인식한다.
plot_num = 1
alg_names = []
eps_list = []
for eps in np.arange(0.1, 1, step=0.1):
alg_names.append("DBScan eps=%.1f" % eps)
eps_list.append(eps)
a = np.arange(len(alg_names) * 2) + 1
indices = np.hstack(a.reshape(len(alg_names), 2).T)
elice_utils.draw_init()
for dataset in datasets:
temp = []
for idx in dataset.index:
# print(str(dataset['1'][idx]) +', '+str(dataset['2'][idx]))
temp.append([dataset["1"][idx], dataset["2"][idx]])
X = np.array(temp)
for alg_name, eps in zip(alg_names, eps_list):
dbscan_result = run_DBScan(X, eps)
elice_utils.draw_graph(X, dbscan_result, alg_name, plot_num, len(alg_names), indices)
plot_num += 1
print(elice_utils.show_graph())
return dbscan_result
def draw_graph(datasets, n_clusters=2):
plot_num = 1
alg_names = []
eps_list = []
for eps in np.arange(0.1, 1, step=0.1):
alg_names.append('DBScan eps=%.1f' % eps)
eps_list.append(eps)
a = np.arange(len(alg_names) * 2) + 1
indices = np.hstack(a.reshape(len(alg_names), 2).T)
elice_utils.draw_init()
for dataset in datasets:
y = dataset['0']
X = np.array(dataset[['1', '2']])
for alg_name, eps in zip(alg_names, eps_list):
dbscan_result = run_DBScan(X, eps)
elice_utils.draw_graph(X, dbscan_result, alg_name, plot_num,
len(alg_names), indices)
plot_num += 1
print(elice_utils.show_graph())
return dbscan_result
def draw_graph(datasets, n_clusters=2, alg_name = 'KMeans'):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
# 3
kmeans_result = run_kmeans(X, n_clusters)
elice_utils.draw_graph(X, kmeans_result, alg_name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return kmeans_result
def draw_graph(datasets, n_clusters=2, alg_name='KMeans'):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
y = dataset['0']
X = np.array(dataset[['1', '2']])
kmeans_result = run_kmeans(X, n_clusters)
elice_utils.draw_graph(X, kmeans_result, alg_name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return kmeans_result
def draw_graph(datasets, n_clusters=2, alg_name = 'KMeans'):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
# 3
y = dataset.ix[:,0]
X = np.vstack((dataset.ix[:,1],dataset.ix[:,2])).T
kmeans_result = run_kmeans(X, n_clusters)
elice_utils.draw_graph(X, kmeans_result, alg_name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return kmeans_result
def draw_graph(datasets, n_clusters=2, eps=0.2, alg_name = ['KMeans', 'DBScan']):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
# 4
X = np.vstack((dataset.ix[:,1],dataset.ix[:,2])).T
kmeans_result = run_kmeans(X, n_clusters)
dbscan_result = run_DBScan(X, eps)
for name, algorithm in zip(alg_name, [kmeans_result, dbscan_result]):
elice_utils.draw_graph(X, algorithm, name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return dbscan_result
def draw_graph(datasets, n_clusters=2, eps=0.2, alg_name=['KMeans', 'DBScan']):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
# 4
X = np.vstack((dataset.ix[:, 1], dataset.ix[:, 2])).T
kmeans_result = run_kmeans(X, n_clusters)
dbscan_result = run_DBScan(X, eps)
for name, algorithm in zip(alg_name, [kmeans_result, dbscan_result]):
elice_utils.draw_graph(X, algorithm, name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return dbscan_result
def draw_graph(datasets, n_clusters=2, alg_name = 'KMeans'):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
# 3
tmp_y = dataset.pop('0')
y = np.array(tmp_y)
tmp_X = dataset
X = np.array(tmp_X)
kmeans_result = run_kmeans(X, n_clusters)
elice_utils.draw_graph(X, kmeans_result, alg_name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return kmeans_result
def draw_graph(datasets, n_clusters=2, alg_name='KMeans'):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
# 3
tmp_y = dataset.pop('0')
y = np.array(tmp_y)
tmp_X = dataset
X = np.array(tmp_X)
kmeans_result = run_kmeans(X, n_clusters)
elice_utils.draw_graph(X, kmeans_result, alg_name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return kmeans_result
def draw_graph(datasets, n_clusters=2, eps=0.2, alg_name = ['KMeans', 'DBScan']):
plot_num = 1
elice_utils.draw_init()
for dataset in datasets:
y = dataset['0']
X = np.array(dataset[['1', '2']])
kmeans_result = run_kmeans(X, n_clusters)
dbscan_result = run_DBScan(X, eps)
for name, algorithm in zip(alg_name, [kmeans_result, dbscan_result]):
elice_utils.draw_graph(X, algorithm, name, plot_num)
plot_num += 1
print(elice_utils.show_graph())
return dbscan_result
