Ruby better than Octave/Matlab?

I’m required to learn GNU Octave or MATLAB. Are they realistically
speaking any better compared to Ruby when it comes to rounding errors
and accuracy?
What about performance?

My context is prototyping machine learning algorithms.


I’m also required to use Matlab. My skills in Ruby are good (I use it
almost every day on a vast variety of tasks), my skills in Matlab are
poor (I use it from time to time).


  • For fast calculations, I use Narray and GSL. I think (but don’t know
    for sure), even than Matlab is faster.
  • With Matlab, I had very bad experience reading and parsing big
    csv-files containing data (it took muuuch longer than ruby).

Rounding errors and accuracy:

  • Don’t know.

I can use Ruby for a lot of different tasks, so I can train myself very
often. Matlab, IMHO, is “only” useful for big calculations, which I do
not need regularly, hence I cannot train it very often and therefore my
skills don’t improve as good as with Ruby.

Kind Regards,

I received answers related to this question from people in the valley.
They say that certain fields, like social science and biology, bet on R,
while engineers bet on Octave/Matlab. None have mentioned Ruby so far.
Comparing Octave and R, you get different results after the 4th decimal
point even in machine learning algorithms. So I guess Ruby is going to
get different results after some decimal points, too. I also was told
that prototyping, people tend to use Octave, NumPy and R before they go
into Java or C++. Whether NArray in Ruby is any worse or better than
Octave’s/Matlab’s highly efficient vector libraries, I’m not sure. It
however would only matters if you have matrices that are close to 10 000
x 10 000, or greater, so I was told.