关联规则 python实现Apriori算法

python实现Apriori算法

根据我们上个博客的例子

def load_dataset():
    # 载入数据集的函数
    dataset = [
        ['A', '', 'D'],
        ['B', 'C', 'E'],
        ['A', 'B', 'C', 'E'],
        ['B', 'E']
    ]
    return dataset

def create_candidates(dataset):
    # 生成1项集的候选集函数
    candidates = []
    for transaction in dataset:
        for item in transaction:
            if [item] not in candidates:
                candidates.append([item])
    candidates.sort()
    return list(map(frozenset, candidates))

def scan_dataset(dataset, candidates, min_support):
    # 扫描数据集，计算候选集的支持度
    item_count = {}  # 记录候选集出现的次数
    for transaction in dataset:
        for candidate in candidates:
            if candidate.issubset(transaction):
                item_count[candidate] = item_count.get(candidate, 0) + 1

    num_transactions = len(dataset)
    frequent_set = []  # 存储频繁项集
    support_data = {}  # 存储支持度数据
    for item in item_count:
        support = item_count[item] / num_transactions
        if support >= min_support:
            frequent_set.append(item)
        support_data[item] = support

    return frequent_set, support_data

def generate_next_candidates(prev_frequent_set, k):
    # 生成下一轮的候选集函数
    next_candidates = []  # 存储下一轮的候选集
    num_frequent_set = len(prev_frequent_set)

    for i in range(num_frequent_set):
        for j in range(i + 1, num_frequent_set):
            item1 = list(prev_frequent_set[i])[:k - 2]
            item2 = list(prev_frequent_set[j])[:k - 2]

            item1.sort()
            item2.sort()

            if item1 == item2:
                next_candidate = prev_frequent_set[i] | prev_frequent_set[j]
                next_candidates.append(next_candidate)

    return next_candidates

def apriori(dataset, min_support=0.5):
    # Apriori 算法主函数
    candidates = create_candidates(dataset)
    dataset = list(map(set, dataset))
    frequent_set1, support_data = scan_dataset(dataset, candidates, min_support)
    frequent_sets = [frequent_set1]
    k = 2

    while len(frequent_sets[k - 2]) > 0:
        candidates = generate_next_candidates(frequent_sets[k - 2], k)
        frequent_set, support_data_k = scan_dataset(dataset, candidates, min_support)
        support_data.update(support_data_k)
        frequent_sets.append(frequent_set)
        k += 1

    return frequent_sets, support_data

# 示例用法
dataset = load_dataset()
frequent_sets, support_data = apriori(dataset, min_support=0.5)

print("频繁项集:")
for i, itemset in enumerate(frequent_sets):
    if itemset:
        print(f"第 {i + 1} 轮: {itemset}")

print("\n支持度数据:")
for item, support in support_data.items():
    print(f"{item}: {support}")

结果

频繁项集:
第 1 轮: [frozenset({‘A’}), frozenset({‘B’}), frozenset({‘C’}), frozenset({‘E’})]
第 2 轮: [frozenset({‘C’, ‘B’}), frozenset({‘E’, ‘B’}), frozenset({‘C’, ‘E’})]
第 3 轮: [frozenset({‘C’, ‘E’, ‘B’})]
支持度数据:
frozenset({‘A’}): 0.5
frozenset({‘D’}): 0.25
frozenset({‘B’}): 0.75
frozenset({‘C’}): 0.5
frozenset({‘E’}): 0.75
frozenset({‘C’, ‘B’}): 0.5
frozenset({‘E’, ‘B’}): 0.75
frozenset({‘C’, ‘E’}): 0.5
frozenset({‘A’, ‘B’}): 0.25
frozenset({‘C’, ‘A’}): 0.25
frozenset({‘A’, ‘E’}): 0.25
frozenset({‘C’, ‘E’, ‘B’}): 0.5