K-Means聚类分析-有标签

模型亮点

1. 初始测试集上评分为0.51，调参后测试集上评分为0.75
2. 数据集由sklearn自带

-----------------------------------------以下为模型具体实现-----------------------------------------

Step1.数据读取

from sklearn.datasets import load_iris
iris=load_iris()
x=iris.data
y=iris.target
import pandas as pd
df_x=pd.DataFrame(x)
df_y=pd.DataFrame(y)
df_x.columns=['sepal_length','sepal_width','petal_length','petal_width']
df_y.columns=['class']

Step2.数据清洗

from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.3,random_state=1)

Step3.启动聚类分析

from sklearn.cluster import KMeans
n=2 #初始参数设置
def kmeans(n,x_train,y_train):
    model=KMeans(n_clusters=n) #初始参数设置
    model.fit(x_train,y_train)
    return model
model=kmeans(n,x_train,y_train)

Step4.模型评估-有标签

from sklearn.metrics import adjusted_rand_score
dic={} #定义空字典，存放兰德指数
def lande(model,x_test,y_test):
    labels_true=y_test
    labels_pred=model.predict(x_test)
    print("兰德指数：",round(adjusted_rand_score(labels_true,labels_pred),2))
    dic[n]=(round(adjusted_rand_score(labels_true,labels_pred),2))
    return model
print("-----初始聚成 2 簇-----")
model=lande(model,x_test,y_test)

Step5.优化参数

for n in range(3,9):
    print("-----聚成",n,"簇-----")
    model=kmeans(n,x_train,y_train) #训练模型
    model=lande(model,x_test,y_test) #评估模型
print("-----最优参数-----")
print("最优簇数：",list(dic.keys())[list(dic.values()).index(max(dic.values()))]) #字典由value反查key
print("最优兰德指数",max(dic.values())) #字典value中最大值

Step6.保存最优模型

n=list(dic.keys())[list(dic.values()).index(max(dic.values()))] #最优簇数
model=kmeans(n,x_train,y_train) #训练最优模型
from sklearn.externals import joblib
joblib.dump(model,'d:\kmeans_labels.pkl')
new_model=joblib.load('d:\kmeans_labels.pkl')
new_model.predict(x_test)

-END