v=mvvacov(x)
#mimetex(D=(\bf{x} - \bf{\mu})^\top \Sigma^{-1} (\bf{x} - \bf{\mu}))
x = fscanfMat("data"); covariance = mvvacov(x); inverse = inv(covariance); for i = 1:3 do centroid(1,i) = mean(x(:,i)); for i = 1:100 do x(i,:) = x(i,:) - centroid(1,:); for i = 1:100 do distance(i) = x(i,:) * inverse * x(i,:)';参考:http://www.ecl.hiroshima-u.ac.jp/~ohno/scilab/man/ja/statistics/mvvacov.htm
http://life.ess.sci.osaka-u.ac.jp/katsura/scilab/index.html
Scilabでデータの書き出し http://sukidesurobocon.blog.shinobi.jp/Entry/291/
参考リンク:Scilab簡易リファレンス, 山本夕可
http://life.ess.sci.osaka-u.ac.jp/katsura/scilab/index.html
deff('[y]=mysin(x)','y=sin(x/180*%pi)') deff('[y]=Rx(x)','y=[1 0 0; 0 cos(x) -sin(x); 0 sin(x) cos(x)]') deff('[y]=Ry(x)','y=[cos(x) 0 -sin(x); 0 1 0; sin(x) 0 cos(x)]') deff('[y]=radian(x)','y=(x/180*%pi)')
http://sukidesurobocon.blog.shinobi.jp/Entry/291/
write('file', arg)