10
$\begingroup$

What modern tools (Windows-based) do you suggest for modeling financial time series?

$\endgroup$
  • $\begingroup$ do you mean a GUI graphical tool that runs on Windows, or a command line based one that runs on Windows (or either) $\endgroup$ – Neil McGuigan Jul 20 '10 at 4:09
  • $\begingroup$ I mean anything (GUI or command line based) that can run on Windows operating systems. $\endgroup$ – Mehper C. Palavuzlar Jul 20 '10 at 6:57
  • $\begingroup$ EViews is a good option for a beginner, but R is a much better long term investment. If you don't have access to it through your Uni, just use R. $\endgroup$ – Jase Dec 16 '12 at 2:47
16
$\begingroup$

I recommend R (see the time series view on CRAN).

Some useful references:

| cite | improve this answer | |
$\endgroup$
7
$\begingroup$

R is great, but I wouldn't really call it "windows based" :) That's like saying the cmd prompt is windows based. I guess it is technically in a window...

RapidMiner is far easier to use [1]. It's a free, open-source, multi-platform, GUI. Here's a video on time series forecasting:

http://rapidminerresources.com/index.php?page=financial-time-series-modelling---part-1

Also, don't forget to read:

http://www.forecastingprinciples.com/

[1] No, I don't work for them.

| cite | improve this answer | |
$\endgroup$
  • 5
    $\begingroup$ There are many R GUI's available for Windows, so I'm not following your point. $\endgroup$ – Shane Jul 19 '10 at 19:34
  • $\begingroup$ I wasn't aware of rapidminer. That looks nice, thanks! $\endgroup$ – James Roth Jul 19 '10 at 20:11
  • 4
    $\begingroup$ "Windows-based" may mean that it works on the Windows operating system (which R does) rather than meaning that it is a heavily GUI-oriented tool. Note the capital letter! $\endgroup$ – seancarmody Jul 20 '10 at 1:01
  • $\begingroup$ fine, you win :) $\endgroup$ – Neil McGuigan Jul 20 '10 at 17:06
  • 1
    $\begingroup$ +1 just for the "No, I don't work for them". Because thats exactly what I thought. :-) $\endgroup$ – vanguard2k Jan 29 '13 at 14:02
4
$\begingroup$

I really like to work with R, because in the end you will find almost anything, and you have a very good support with the mailing lists. The downside of R is that helpful bits which fit your specific problems might be spread over a large range of packages, and you might not always be able to find them. Another point may be a lock-in, with that I mean that after a time learning R, you will probably be unmotivated to relearn another software, but this will happen in any system.

With regard to Matlab being expensive - if on a budget, Octave will work just as well, at least it did for the things I needed to do with it, which were rather basic.

$\endgroup$
2
$\begingroup$

I'm new here, and perhaps "financial time series" has a specific definition... But given that I don't know it, my question for you would be what you mean: quarterly/monthly economic data, daily market prices, hourly or higher-frequency data, etc? And by "modeling", do you mean working with textbook ARIMA/ARCH solutions, or things a bit more exotic (such as dynamic linear systems), or exotic/custom experimentation?

R is flexible and free, though less GUI-fied than most. It also has packages covering everything from daily stock prices to dynamic linear systems and optimization packages. (In fact, the hard part will be deciding which time series and which financial packages to use.)

GRETL is free and has a reasonable GUI, though it's econometric, not really daily market oriented. I've heard of Oxmetrics, which appears to have a very complete every-possible-variant-of-ARCH package available for it. If you're talking monthly/quarterly economic data, you could also use X12-ARIMA, which is a benchmark of sorts.

I've used all kinds of GUIs for programming/processing data, but for some reason RapidMiner's never really clicked with me. Something strange about its workflow that I've just never gotten.

| cite | improve this answer | |
$\endgroup$
2
$\begingroup$

While not exactly cheap, MATLAB is widely used in the financial industry for time series modelling: http://www.mathworks.com

$\endgroup$
1
$\begingroup$
  • Clearly R

  • RadidMiner is nice, but switching to thinking in terms of operators takes a moment

  • Matlab / Octave

If you describe a specific problem, I may be able to get more specific.

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

At my university, Stata is taught as a programme to do statistical analysis for finance. You can use outreg for example to format tables for publications in financial papers very easily. Programming syntax is not really great I think, you have to declare functions with `variable' for example which is a quirk in my opinion. Amount of different statistical functions however is very vast.

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

Probably not exactly what you are looking for, but you may check SwiftForecast. It allows you to forecast a time series in an automatic way, without the need of any software. It is quite new, but I find the idea of a "Google style" predictor quite interesting...

$\endgroup$
0
$\begingroup$

You might want to consider using LDT. It is free and while it provides automatic forecasting with stationary vector autoregressive (VAR) models, you can benefit form other types of analysis.

PS: I am the developer of this software.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.