Machine [unsupervised{Clustering(Hierarchical)}] (계층적, 병합적 군집)agglomerative

Hierarchical Clustering¶

Library 임포트¶

In [1]:

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

In [ ]:

In [ ]:

dataset 읽어오기¶

In [3]:

df = pd.read_csv("../data/Mall_Customers.csv")

In [4]:

df

Out[4]:

	CustomerID	Genre	Age	Annual Income (k$)	Spending Score (1-100)
0	1	Male	19	15	39
1	2	Male	21	15	81
2	3	Female	20	16	6
3	4	Female	23	16	77
4	5	Female	31	17	40
...	...	...	...	...	...
195	196	Female	35	120	79
196	197	Female	45	126	28
197	198	Male	32	126	74
198	199	Male	32	137	18
199	200	Male	30	137	83

200 rows × 5 columns

In [ ]:

In [ ]:

In [ ]:

Training the Hierarchical Clustering model¶

In [5]:

X = df.loc[:, 'Annual Income (k$)':]

In [6]:

X

Out[6]:

	Annual Income (k$)	Spending Score (1-100)
0	15	39
1	15	81
2	16	6
3	16	77
4	17	40
...	...	...
195	120	79
196	126	28
197	126	74
198	137	18
199	137	83

200 rows × 2 columns

In [ ]:

Dendrogram 을 그리고, 최적의 클러스터 갯수를 찾아보자.¶

In [7]:

import scipy.cluster.hierarchy as sch

In [12]:

sch.dendrogram( sch.linkage(X, method = 'ward') )  # ward 방식으로 묶는다.
plt.title("Dendrogram")
plt.xlabel("'Customers'")
plt.ylabel("Euclidean Distances")
plt.show

Out[12]:

<function matplotlib.pyplot.show(close=None, block=None)>

In [ ]:

Training the Hierarchical Clustering model¶

In [13]:

from sklearn.cluster import AgglomerativeClustering

In [14]:

hc = AgglomerativeClustering(n_clusters= 5)

In [16]:

y_pred = hc.fit_predict(X)

In [17]:

df["Group"] = y_pred

In [18]:

df

Out[18]:

	CustomerID	Genre	Age	Annual Income (k$)	Spending Score (1-100)	Group
0	1	Male	19	15	39	4
1	2	Male	21	15	81	3
2	3	Female	20	16	6	4
3	4	Female	23	16	77	3
4	5	Female	31	17	40	4
...	...	...	...	...	...	...
195	196	Female	35	120	79	2
196	197	Female	45	126	28	0
197	198	Male	32	126	74	2
198	199	Male	32	137	18	0
199	200	Male	30	137	83	2

200 rows × 6 columns

In [ ]:

In [ ]:

그루핑 정보를 확인¶

In [22]:

plt.figure(figsize=[12,8])
plt.scatter(X.values[y_pred == 0, 0], X.values[y_pred == 0, 1], s = 100, c = 'red', label = 'Cluster 1')
plt.scatter(X.values[y_pred == 1, 0], X.values[y_pred == 1, 1], s = 100, c = 'blue', label = 'Cluster 2')
plt.scatter(X.values[y_pred == 2, 0], X.values[y_pred == 2, 1], s = 100, c = 'green', label = 'Cluster 3')
plt.scatter(X.values[y_pred == 3, 0], X.values[y_pred == 3, 1], s = 100, c = 'cyan', label = 'Cluster 4')
plt.scatter(X.values[y_pred == 4, 0], X.values[y_pred == 4, 1], s = 100, c = 'magenta', label = 'Cluster 5')
plt.title('Clusters of customers')
plt.xlabel('Annual Income (k$)')
plt.ylabel('Spending Score (1-100)')
plt.legend()
plt.show()

In [ ]:

In [ ]:

그룹 정보를 데이터셋에 추가¶

In [ ]:

In [ ]:

In [ ]:

마케팅 이메일을 보내기 위해, 3번 그룹의 고객들만 가져와보자.¶

In [23]:

df.loc[df["Group"]==3, ] 

Out[23]:

	CustomerID	Genre	Age	Annual Income (k$)	Spending Score (1-100)	Group
1	2	Male	21	15	81	3
3	4	Female	23	16	77	3
5	6	Female	22	17	76	3
7	8	Female	23	18	94	3
9	10	Female	30	19	72	3
11	12	Female	35	19	99	3
13	14	Female	24	20	77	3
15	16	Male	22	20	79	3
17	18	Male	20	21	66	3
19	20	Female	35	23	98	3
21	22	Male	25	24	73	3
23	24	Male	31	25	73	3
25	26	Male	29	28	82	3
27	28	Male	35	28	61	3
29	30	Female	23	29	87	3
31	32	Female	21	30	73	3
33	34	Male	18	33	92	3
35	36	Female	21	33	81	3
37	38	Female	30	34	73	3
39	40	Female	20	37	75	3
41	42	Male	24	38	92	3

In [ ]:

In [ ]:

# 각 그룹별 수입과 소비지표의 평균을 구하세요.

In [27]:

pd.pivot_table(df, index="Group", aggfunc = np.mean)

Out[27]:

	Age	Annual Income (k$)	CustomerID	Spending Score (1-100)
Group
0	41.000000	89.406250	166.250000	15.593750
1	42.482353	55.811765	87.894118	49.129412
2	32.692308	86.538462	162.000000	82.128205
3	25.333333	25.095238	22.000000	80.047619
4	45.217391	26.304348	23.000000	20.913043

In [26]:

df.groupby("Group")[['Annual Income (k$)','Spending Score (1-100)']].describe()

Out[26]:

	Annual Income (k$)								Spending Score (1-100)
	count	mean	std	min	25%	50%	75%	max	count	mean	std	min	25%	50%	75%	max
Group
0	32.0	89.406250	16.612975	71.0	78.0	86.5	98.25	137.0	32.0	15.593750	8.936548	1.0	9.75	15.0	20.5	39.0
1	85.0	55.811765	9.731508	39.0	48.0	57.0	63.00	79.0	85.0	49.129412	7.281399	29.0	43.00	49.0	55.0	65.0
2	39.0	86.538462	16.312485	69.0	75.5	79.0	95.00	137.0	39.0	82.128205	9.364489	63.0	74.50	83.0	90.0	97.0
3	21.0	25.095238	7.133756	15.0	19.0	24.0	30.00	38.0	21.0	80.047619	10.249274	61.0	73.00	77.0	87.0	99.0
4	23.0	26.304348	7.893811	15.0	19.5	25.0	33.00	39.0	23.0	20.913043	13.017167	3.0	9.50	17.0	33.5	40.0