数据挖掘 | 数据分析师都在看的聚类分析知识点总汇

cloumns_fix1 = ['工作日上班时电话时长', '工作日下半时电话时长',     '周末电话时长', '国际电话时长', '平均每次通话时长']
#数据降维pca_2 = PCA(n_components=2)data_pca_2 = pd.DataFrame(pca_2.fit_transform(data[cloumns_fix1]))

#绘制散点图查看数据点大致情况plt.scatter(data_pca_2[0],data_pca_2[1])

#预计将数据点分类为3类kmmodel = KMeans(n_clusters=3) #创建模型kmmodel = kmmodel.fit(data[cloumns_fix1]) #训练模型ptarget = kmmodel.predict(data[cloumns_fix1]) #对原始数据进行标注
pd.crosstab(ptarget,ptarget) #交叉表查看各个类别数据的数量

plt.scatter(data_pca_2[0],data_pca_2[1],c=ptarget)#查看聚类的分布情况

#查看各类之间的差异dMean = pd.DataFrame(columns=cloumns_fix1+['分类']) #得到每个类别的均值data_gb = data[cloumns_fix1].groupby(ptarget) #按标注进行分组
i = 0for g in data_gb.groups:         rMean = data_gb.get_group(g).mean()         rMean['分类'] = g;         dMean = dMean.append(rMean, ignore_index=True)         subData = data_gb.get_group(g)                for column in cloumns_fix1:                 i = i+1;                 p = plt.subplot(3, 5, i)                 p.set_title(column)                 p.set_ylabel(str(g) + "分类")                 plt.hist(subData[column], bins=20)

cloumns_fix1 = ['工作日上班时电话时长', '工作日下半时电话时长',     '周末电话时长',          '国际电话时长', '平均每次通话时长']          linkage = hcluster.linkage(data[cloumns_fix1], method='centroid') #中心点距离计算，得到矩阵

#层次聚类绘图hcluster.dendrogram(linkage)  #不设置参数时会将所有点做为一个基础的类进行树结构的绘制#由于数据量大，限制类的个数,保留12个节点，有括号表示副节点，括号内的数字为该节点内部包含的子节点hcluster.dendrogram(linkage, truncate_mode='lastp', p=12, leaf_font_size=12.)

#对聚类得到的类进行标注       层次聚类的结果，要聚类的个数，划分方法（maxclust，最大划分法）ptarget = hcluster.fcluster(linkage, 3, criterion='maxclust')#查看各类别中样本含量pd.crosstab(ptarget,ptarget)

#使用主成分分析进行数据降维pca_2 = PCA(n_components=2)data_pca_2 = pd.DataFrame(pca_2.fit_transform(data[cloumns_fix1]))
plt.scatter(data_pca_2[0], data_pca_2[1], c=ptarget) #绘制图形

#计算得到各点间距离的矩阵from sklearn.metrics.pairwise import euclidean_distancesdist = euclidean_distances(data)

#设置Eps和MinPtseps = 0.2MinPts = 5
ptses = []for row in dist:   #密度    density = np.sum(row<eps)     pts = 0        if density>MinPts:  #核心点,密度大于5              pts = 1        elif density>1 :   #边界点，密度大于1小于5        pts = 2       else:    #噪声点，密度为1        pts = 0    ptses.append(pts)#得到每个点的分类

#把噪声点过滤掉，因为噪声点无法聚类，它们独自一类corePoints = data[pandas.Series(ptses)!=0]coreDist = euclidean_distances(corePoints)

cluster = dict()i = 0for row in coreDist:     cluster[i] = numpy.where(row<eps)[0]    i = i + 1

for i in range(len(cluster)):    for j in range(len(cluster)):        if len(set(cluster[j]) & set(cluster[i]))>0 and i!=j:            cluster[i] = list(set(cluster[i]) | set(cluster[j]))            cluster[j] = list()

result = dict()j = 0for i in range(len(cluster)):  if len(cluster[i])>0:    result[j] = cluster[i]    j = j + 1#找出每个点所在领域的序号，作为他们最后聚类的结果标记for i in range(len(result)):    for j in result[i]:        data.at[j, 'type'] = i plt.scatter(data['x'], data['y'], c=data['type'])

eps = 0.2MinPts = 5
model = DBSCAN(eps, MinPts)
data['type'] = model.fit_predict(data)
plt.scatter(data['x'],  data['y'], c=data['type'])