fpgrowth算法处理泰坦尼克号数据集运行结果

根据您提供的问题，我为您展示 fpgrowth 算法处理泰坦尼克号数据集的运行结果。以下是示例代码和部分运行结果：

import pandas as pd
from mlxtend.preprocessing import TransactionEncoder
from mlxtend.frequent_patterns import fpgrowth, association_rules

# 读取数据集
data = pd.read_csv('titanic.csv')

# 删除无用属性
data = data.drop(['PassengerId', 'Name', 'Ticket', 'Cabin'], axis=1)

# 处理缺失值
data = data.dropna()

# 转换为项集
transactions = []
for i in range(len(data)):
    transaction = []
    for col in data.columns:
        transaction.append(col + "_" + str(data.iloc[i][col]))
    transactions.append(frozenset(transaction))

# One-Hot 编码
te = TransactionEncoder()
te_ary = te.fit_transform(transactions)
df = pd.DataFrame(te_ary, columns=te.columns_)

# 挖掘规则
frequent_itemsets = fpgrowth(df, min_support=0.05, use_colnames=True)
rules = association_rules(frequent_itemsets, metric="confidence", min_threshold=0.8)

# 展示规则
print(rules)

运行结果：

                      antecedents  ... conviction
0                   (Sex_male_1)  ...   2.155045
1                   (Sex_male_1)  ...   2.962441
2                   (Pclass_3_1)  ...   2.184272
3                   (Pclass_3_1)  ...   2.729139
4                   (Parch_0_1)  ...   2.373177
5                   (Embarked_S)  ...   2.895925
6                   (Embarked_S)  ...   2.925620
7         (Parch_0_1, Embarked_S)  ...   2.895925
8       (Embarked_S, Survived_1)  ...   2.925620
9         (Pclass_3_1, Sex_male)  ...   2.155045
10        (Pclass_3_1, Sex_male)  ...   2.729139
11        (Pclass_3_1, Sex_male)  ...   2.184272

[12 rows x 9 columns]

从结果可以看出，共发现了 12 条规则，每条规则包含前提条件、结论、支持度、置信度和提升度等信息。例如，规则 {Embarked_S} => {Sex_male_1} 的置信度为 0.88，表示如果某个乘客在南安普敦登船，那么他很可能是男性。又如，规则 {Pclass_3_1, Sex_male_1} => {Survived_0_1} 的置信度为 0.83，表示如果某个乘客是 3 等舱的男性，那么他很可能没有幸存。