We have an N*N matrix, can we achieve better than transposing the matrix into an array, and then sorting it?
I did some online searching and found that quickselect eliminates the log factor, how can we extend quickselect to 2 dimensions?
# | User | Rating |
---|---|---|
1 | tourist | 3803 |
2 | jiangly | 3707 |
3 | Benq | 3627 |
4 | ecnerwala | 3584 |
5 | orzdevinwang | 3573 |
6 | Geothermal | 3569 |
6 | cnnfls_csy | 3569 |
8 | Radewoosh | 3542 |
9 | jqdai0815 | 3532 |
10 | gyh20 | 3447 |
# | User | Contrib. |
---|---|---|
1 | awoo | 163 |
1 | maomao90 | 163 |
3 | adamant | 161 |
4 | maroonrk | 152 |
5 | -is-this-fft- | 151 |
6 | nor | 150 |
7 | atcoder_official | 147 |
7 | SecondThread | 147 |
9 | TheScrasse | 146 |
10 | Petr | 145 |
We have an N*N matrix, can we achieve better than transposing the matrix into an array, and then sorting it?
I did some online searching and found that quickselect eliminates the log factor, how can we extend quickselect to 2 dimensions?
Name |
---|
Convert the matrix to an array first, then use
nth_element(x,x+n/2,x+n)
. The overall complexity is the same as the input complexity, which has reached the lower bound.nth_element is $$$\Theta(n)$$$.
nth_element is $$$O(n) $$$, not $$$\Theta(n)$$$
You're wrong, it's $$$\Theta(n)$$$, because it takes at least a linear perusal of all the elements to determine the value of the nth.
Just use the Thanos algorithm: among $$$n^2$$$ elements of the matrix, select random $$$n$$$ and find their median with
nth_element
;-)How can we efficiently select n random elements from the matrix without iterating over the full matrix?
And let's we want to extend the program, and find the smallest median among every submatrix of size K*K, how can we solve this?
binary search
With binary search wouldnt it be N^2*K*logK, but can we achieve N^2*K or better?
Auto comment: topic has been updated by MODDI (previous revision, new revision, compare).