Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
As an example, If I had inputs of data that are then analyzed by Rapidminer I would take the analysis of that and need to input it into a Ninja strategy based on that analysis. What is the conduit between Ninja and Rapidminer? How does the information transfer?
Signed, Clueless in Programming/Scripting
Jungian
Can you help answer these questions from other members on NexusFi?
I dabbled with this for a while and used a lot of different data mining softwares as well as some homegrown versions. What you're going to need to do for this is to use the .dll files that are included with RapidMiner or whatever you do end up using. From there you will need to reference the saved model/network that you have trained.
It's probably best to use Visual Studio for this so that you can see what the .dll file references and how to access it. After you can access your network though, you should be fine. It's just a matter of taking the indicators and other inputs you trained it to use and plugging them in.
The best suites I used and the ones I had the most luck implementing in NinjaTrader were:
with AForge being an order of magnitude faster than Encog, although Encog is more fully-featured. At least in Java, I don't know if the C# library has caught up yet. Implementation between the two was almost exactly the same. If you need more guidance than that, I probably have some legacy code lying around for how to hook up to each one, but it'll take some searching.
Finally, I did notice you said "data mining" and a word of advice; I went through the same thing and what you should really be looking into is "time-series analysis", which is a whole different kind of analysis. Instead of looking at Neural Nets, Perceptrons, and Feedforward Backpropagation this or that, you should be looking at ANOVA and MANOVA(spelling?). However, if you do use Encog, I believe the have the Boltzmann machine is now a part of the library and I remember really wanting to look into that after looking at a few talks on the subject. I think there is a lot of promise there.
I derived this from 6 years of 5min candles. Using Access for bulk manipulation and Excel for ad-hoc reporting and charting. Some interesting results...