（Apriori）とアソシエーションルール

# Program Name: market_basket_analysis.py  
# Creation Date: 20250606  
# Overview: Perform market basket analysis using Apriori and visualize association rules  
# Usage: Run in Python (Jupyter/Colab) with mlxtend, matplotlib, pandas installed

# -------------------- Install Required Libraries --------------------
!pip install pandas mlxtend matplotlib seaborn --quiet

# -------------------- Import Libraries --------------------
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from mlxtend.preprocessing import TransactionEncoder
from mlxtend.frequent_patterns import apriori, association_rules

# -------------------- Parameters / 取引データ --------------------
# 仮想的な顧客購入履歴 / Sample transactions
dataset = [
    ['milk', 'bread', 'butter'],
    ['bread', 'diaper', 'beer', 'eggs'],
    ['milk', 'bread', 'diaper', 'beer'],
    ['milk', 'bread', 'diaper', 'butter'],
    ['bread', 'butter']
]

# -------------------- Step 1: データのバイナリ化 / Encode the dataset --------------------
te = TransactionEncoder()
te_ary = te.fit(dataset).transform(dataset)
df = pd.DataFrame(te_ary, columns=te.columns_)

# -------------------- Step 2: 頻出アイテムセット抽出 / Extract Frequent Itemsets --------------------
frequent_itemsets = apriori(df, min_support=0.4, use_colnames=True)

# -------------------- Step 3: アソシエーションルールの抽出 / Derive Association Rules --------------------
rules = association_rules(frequent_itemsets, metric="lift", min_threshold=1.0)

# -------------------- Step 4: ルール上位表示 / Display Top Rules --------------------
print(rules[['antecedents', 'consequents', 'support', 'confidence', 'lift']].sort_values(by='lift', ascending=False))

# -------------------- Step 5: プロット（Lift vs Confidence） --------------------
plt.figure(figsize=(8, 6))
sns.scatterplot(data=rules, x='confidence', y='lift', size='support', hue='support', palette='viridis', sizes=(100, 1000), legend=False)

# 英語表記 / Plot labels in English
plt.title('Association Rules: Lift vs Confidence')
plt.xlabel('Confidence')
plt.ylabel('Lift')
plt.grid(True)
plt.show()