我有一个 user:friends (50,000) 列表和一个事件参与者列表(25,000 个事件和每个事件的参与者列表)。我想找到与用户一起参加事件的前 k 个 friend 。这需要为每个用户完成。
我试过遍历列表，但计算量很大。我也在尝试通过创建加权图来做到这一点。(Python)
如果有任何其他方法，请告诉我。
请您参考如下方法:
Python 的集合对象(字典、集合和 collections.Counter)简化了这项任务:
from collections import Counter
def top_k_friends(friends, events, k=2):
'''Given a dictionary users mapped to their set of friends
and a dictionary of events mapped to a set of their attendees,
find the top k friends with whom the user goes to the event.
Do this for each user.
'''
for user, users_friends in friends.iteritems():
c = Counter()
for event, attendees in events.iteritems():
if user in attendees:
c.update(users_friends.intersection(attendees))
print user, '-->', c.most_common(k)
if __name__ == '__main__':
friends = {
'robert' : {'mary', 'marty', 'maggie', 'john'},
'paul' : {'marty', 'mary', 'amber', 'susan'}
}
events = {
'derby': {'amber', 'mary', 'robert'},
'pageant': {'maggie', 'paul', 'amber', 'marty', 'john'},
'fireworks': {'susan', 'robert', 'marty', 'paul', 'robert'}
}
top_k_friends(friends, events)
