aravind.nujella's blog

Find users with highest rating change in an interval
By aravind.nujella, history, 4 hours ago, In English

Assume a codeforces like platform where there are $$$n$$$ users and their ratings vary over time $$$[0, T]$$$. Now given an interval $$$[l,r]$$$, we would like to find top k profiles which had highest delta in that interval.

Is it possible to answer this query without iterating over $$$n$$$ interval sums? For example we can get $$$O(n\log(n)\log(T))$$$ by maintaing fenwick tree per user and sorting the list of interval sums.

Edit: Added requirement to avoid looping over $$$n$$$ users..

  • Vote: I like it
  • 0
  • Vote: I do not like it

»
3 hours ago, # |
← Rev. 2   Vote: I like it 0 Vote: I do not like it

Cant you just use prefix sums of each profile. Then use PriorityQueue and in total you need to precompute in O(T*N) and each query works for O(N+K).

→ Reply
  • »
    »
    3 hours ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Thats true. What I should have asked for is to reduce $$$n$$$ factor in the query since I do not want to iterate over list of users.

    → Reply
    • »
      »
      »
      3 hours ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      You can store it in dp[l,r] and do query in O(K) but your space complexity will be very very large. I think there is more smart way, but its first what comes to mind.

      → Reply
      • »
        »
        »
        »
        3 hours ago, # ^ |
          Vote: I like it 0 Vote: I do not like it

        Yeah. That DP is not feasible because $$$T$$$ can be quite big.

        → Reply
»
3 hours ago, # |
  Vote: I like it 0 Vote: I do not like it

Auto comment: topic has been updated by aravind.nujella (previous revision, new revision, compare).

→ Reply