Obsidian-Knowledge-Base/X-Knowledge-Base/raw/Joplin/_resources/instacart_pca.py

#!/usr/bin/env python
# coding: utf-8

# In[24]:


# 1、获取数据
# 2、合并表
# 3、找到user_id和aisle之间的关系
# 4、PCA降维


# In[25]:


import pandas as pd


# In[26]:


# 1、获取数据
order_products = pd.read_csv("../main_01/instacart/order_products__prior.csv")
products = pd.read_csv("../main_01//instacart/products.csv")
orders = pd.read_csv("../main_01//instacart/orders.csv")
aisles = pd.read_csv("../main_01//instacart/aisles.csv")


# In[27]:


# 2、合并表
# order_products__prior.csv：订单与商品信息

# 字段：order_id, product_id, add_to_cart_order, reordered
# products.csv：商品信息
# 字段：product_id, product_name, aisle_id, department_id
# orders.csv：用户的订单信息
# 字段：order_id,user_id,eval_set,order_number,….
# aisles.csv：商品所属具体物品类别
# 字段： aisle_id, aisle

# 合并aisles和products aisle和product_id
tab1 = pd.merge(aisles, products, on=["aisle_id", "aisle_id"])


# In[28]:


tab2 = pd.merge(tab1, order_products, on=["product_id", "product_id"])


# In[29]:


tab3 = pd.merge(tab2, orders, on=["order_id", "order_id"])


# In[30]:


tab3.head()


# In[31]:


# 3、找到user_id和aisle之间的关系
table = pd.crosstab(tab3["user_id"], tab3["aisle"])


# In[32]:


data = table[:10000]


# In[33]:


# 4、PCA降维
from sklearn.decomposition import PCA


# In[34]:


# 1）实例化一个转换器类
transfer = PCA(n_components=0.95)

# 2）调用fit_transform
data_new = transfer.fit_transform(data)


# In[35]:


data_new.shape


# In[36]:


data_new


# In[37]:


# 预估器流程
from sklearn.cluster import KMeans


# In[38]:


estimator = KMeans(n_clusters=3)
estimator.fit(data_new)


# In[39]:


y_predict = estimator.predict(data_new)


# In[40]:


y_predict[:300]


# In[41]:


# 模型评估-轮廓系数
from sklearn.metrics import silhouette_score


# In[42]:


silhouette_score(data_new, y_predict)