1
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 1 year has passed since last update.

僕も Ruby と Python で pagepank をやってみた

Last updated at Posted at 2022-02-05

はじめに

上記URLにあります、PythonのコードをRubyでやってみようと思います。

result.py
M = np.array([[0, 0, 0, 0, 1],
              [0.5, 0, 0, 0, 0],
              [0.5, 0, 0, 0, 0],
              [0, 1, 0.5, 0, 0],
              [0, 0, 0.5, 1, 0]])
v = pagerank(M, 100, 1)
print(v)

# output
[[0.26658293] 
 [0.13322816] 
 [0.13322816] 
 [0.20009761] 
 [0.26686313]]

ここでは、ダンピング・ファクターを1とします。

問題の整理(AtCoder風)

ノード数Nの有向グラフがある。
高橋くんはノード1からスタートして100回ランダムな遷移を行います。
遷移後、高橋くんがどのノードにいるかを答えてください。

入力
N M
A1 B1
.
.
.
Am Bm
入力例
5 7
1 2
1 3
2 4
3 4
3 5
4 5
5 1

Ruby

n, m = gets.split.map(&:to_i)
ab = Array.new(n){ Array.new(n, 0) }
m.times do
  a, b = gets.split.map(&:to_i)
  ab[a - 1][b - 1] = 1
end
ab.each_with_index do |x, i|
  s = x.sum
  x.each_with_index do |_, j|
    ab[i][j] /= s.to_f
  end
end
m_hat = ab.transpose

v = n.times.map{ rand(0...1.0) }
vs = v.sum
v.map!{ _1 / vs }
100.times do
  vd = v.dup
  n.times do |i|
    v[i] = 0
    n.times do |j|
      v[i] += m_hat[i][j] * vd[j]
    end
  end
end
puts v

# output
0.2662026143788123
0.13329803377108196
0.13329803377108196
0.20040578221512312
0.2667955358639006

ふう、TLEにならなければよいのですが。

[追記] Python

import networkx as nx

N, M = map(int, input().split())
G = nx.DiGraph()
for _ in range(M):
    A, B = map(int, input().split())
    nx.add_path(G, [A, B])

pr = nx.pagerank(G, alpha=0.95, max_iter=100)
print(pr)

# output
{1: 0.26261033194706523, 2: 0.13473925873325307, 3: 0.13473925873325307, 4: 0.20200447126193546, 5: 0.26590667932449313}

いやー、python便利、networkx便利ですね。
ちなみに、ダンピング・ファクターであるalphaは0.95を超えるとエラーになる様です。

まとめ

  • pegerankに詳しくなった
1
0
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
1
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?