Naive implementation of the Apriori algorithm in Python

apriori.py

# Naive implementation of the Apriori algorithm in Python

# Example 2 from https://en.wikipedia.org/wiki/Apriori_algorithm
data = [
    {1,2,3,4},
    {1,2,4},
    {1,2},
    {2,3,4},
    {2,3},
    {3,4},
    {2,4}
]

def support(tx, txs, absolute=False):
    if absolute: return len([t for t in txs if t.issuperset(tx)])
    return len([t for t in txs if t.issuperset(tx)]) / len(txs)

def apriori(txs, min_sup):
    l = [[{i} for i in set().union(*txs) if support({i}, txs, min_sup > 1) > min_sup]]
    k = 1
    while len(l[k-1]) > 0:
        c = list(set([frozenset(s.union(i)) for i in l[0] for s in l[k-1] if s != i and len(i) == 1 and len(s.union(i)) == k+1])) # frozensets are hashable
        l.insert(k, [ct for ct in c if support(ct, txs, min_sup > 1) >= min_sup])
        k += 1
    return l
    
if __name__ == '__main__':
    min_sup = 3
    fis = apriori(data, min_sup)
    if len(fis[0]) < 1: print('Did not find any frequent item sets with support of {}'.format(min_sup))
    else: print('Maximum length ({}) frequent item sets are: {}'.format(len(fis[-2][0]), [tuple(s) for s in fis[-2]]))