Skip to content

Instantly share code, notes, and snippets.

@kiyotune
Last active December 11, 2015 15:08
Show Gist options
  • Save kiyotune/4618451 to your computer and use it in GitHub Desktop.
Save kiyotune/4618451 to your computer and use it in GitHub Desktop.
https://codeiq.jp/ace/kurahashi_issei/q166 の回答。 ただしOctaveによる。 だってR言語知らないもん。
#!/usr/local/bin/octave -q
1;
% disp_stat:
function disp_stat(n, height=[])
if length(height) == 0 && n > 0
height = 170 + 10 * randn(1, n);
end
printf("%5d\t%9.5f\t%9.5f\t%9.5f\t%9.5f\n",
length(height),
var(height, 1),
q2 = var(height, 0),
std(height, 1),
sqrt(q2) / length(height)
);
endfunction
% Q1.
disp('<Q1>');
height = [168, 173, 152, 181, 175];
printf("[平均値, 中央値] = [%f, %f]\n",
mean(height),
median(height)
);
printf("N\t標本分散\t不偏分散\t標準偏差\t標準誤差\n");
disp_stat(-1, height);
% Q2.
disp('<Q2>');
printf("N\t標本分散\t不偏分散\t標準偏差\t標準誤差\n");
disp_stat(5);
disp_stat(100);
disp_stat(1000);
disp_stat(10000);
% ans
% <Q1>
% [平均値, 中央値] = [169.800000, 173.000000]
% N 標本分散 不偏分散 標準偏差 標準誤差
% 5 96.56000 120.70000 9.82649 2.19727
% <Q2>
% N 標本分散 不偏分散 標準偏差 標準誤差
% 5 45.55437 56.94297 6.74940 1.50921
% 100 76.49337 77.26603 8.74605 0.08790
% 1000 103.39394 103.49744 10.16828 0.01017
% 10000 99.16559 99.17551 9.95819 0.00100
%
% サンプルの数(N)が増えるにつれ、
% ・標本分散≒不偏分散の傾向
% ・標準偏差は10付近に収束傾向
% ・標準誤差は0に収束傾向
% がある。
%
% => だから何なのかはわかりませんでした。
@kiyotune
Copy link
Author

なんで"Objective-C"と判定されとるとですか...

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment