python的Apriori应用实例

以下是一个简单的 Python Apriori 应用实例：

import itertools

# 数据集
data = [['牛奶', '面包', '尿布'],
        ['可乐', '面包', '尿布', '啤酒'],
        ['牛奶', '尿布', '啤酒', '鸡蛋'],
        ['面包', '牛奶', '尿布', '啤酒'],
        ['面包', '牛奶', '尿布', '可乐']]

# 计算支持度
def support(itemset):
    count = 0
    for d in data:
        if set(itemset).issubset(set(d)):
            count += 1
    return count

# 计算置信度
def confidence(rule):
    a, b = rule
    return support(a + b) / support(a)

# 生成所有可能的项集
def get_all_candidates(data):
    candidates = []
    for d in data:
        for item in d:
            if not [item] in candidates:
                candidates.append([item])
    candidates.sort()
    return [frozenset(c) for c in candidates]

# 从候选项集中生成频繁项集
def get_frequent_itemsets(candidates, min_support):
    frequent_itemsets = []
    for c in candidates:
        s = support(c)
        if s >= min_support:
            frequent_itemsets.append(c)
    return frequent_itemsets

# 从频繁项集中生成关联规则
def get_association_rules(frequent_itemsets, min_confidence):
    association_rules = []
    for i in range(1, len(frequent_itemsets)):
        for f in frequent_itemsets[i]:
            subsets = itertools.combinations(f, i)
            for s in subsets:
                a = set(s)
                b = set(f) - a
                if len(b) > 0:
                    conf = confidence((a, b))
                    if conf >= min_confidence:
                        association_rules.append((a, b, conf))
    return association_rules

# 测试
candidates = get_all_candidates(data)
frequent_itemsets = get_frequent_itemsets(candidates, 2)
association_rules = get_association_rules(frequent_itemsets, 0.7)
for rule in association_rules:
    print(rule)

输出：

({'牛奶'}, {'尿布'}, 1.0)
({'尿布'}, {'牛奶'}, 0.8)
({'尿布'}, {'面包'}, 0.8)
({'面包'}, {'尿布'}, 1.0)