October 13, 2008
Rapid-I Meeting
I had the pleasure of meeting Ralf Klinkenberg, founder of Rapid-I, this weekend in New York City. He was in town finishing up the first set of Rapidminer training courses and he dropped me an email to see if I had some time to meet. We talked about the birth of Rapidminer (formerly Y.A,L.E.), why he started it, what’s the ultimate plan for Rapidminer, and their growing team in Dortmund.
We had a good long talk over Malaysian food and decided to strengthen ties between Rapid-I and Neural Market Trends. We even talked about the possibility of me co-presenting one of their financial training courses next year. I must say, I walked away rather excited about the possibilities and I look forward to growing our relationship in the future!
Oh BTW, Ralf will be back next week in NYC for more Rapidminer training courses. The courses he’s presenting focus on an introduction to datamining and financial datamining and analysis. I’ll post the exact list of courses and availability on here later, so check back!


October 14th, 2008 at 6:01 pm
Hi Tom,
Very cool…
I hope this means we will see some more tutorials from you soon dude ;-)
Also like the new format of your site… Its on topic 8^)
Keep up the good work!!!
Cordially,
-Digital Dude-
“You’re only as good as your next picture.” -Walt Disney-
October 14th, 2008 at 8:16 pm
Tom – I attended the Rapid Miner seminars this past week in NYC. They were very beneficial.
Ralf & I had talked about adding some sort of financial time series pre-processing operators to reduce the work that would need to be performed outside of Rapid Miner prior to building a model.
Two existing java libraries for this functionality came to mind – TA-Lib & QuantLib. I think TA-Lib may be more useful & easier for most users to use & understand.
Any thoughts on adding these operators? I know you have more experience with the product than I do so any insight you have would be great. I haven’t fully thought out what all I think would be useful to have, but I am sure we can make generate some good ideas for the product.
Feel free to email me to discuss further.
Regards,
Eric
October 15th, 2008 at 4:36 am
DD: I’ll be posting videos soon again. Thanks for being patient!
Eric: Are you referring to Mathlab libraries? I’ve been thinking about this for a while myself but I was leaning more to a database driven query system, but something like this would be vastly easier. Unfortunately, I very little Mathlib experience from my Freshman year in Engineering. That was a long time ago.
October 15th, 2008 at 7:52 am
Tom- I do not use MatLab so I am unsure of its functionality. I would like to learn more about it but havent had the chance. Also, I believe the license is fairly pricy to play around with.
I am referring to TA-Lib as a set of technical analysis pre-processing operators (http://www.ta-lib.org) and/or QuantLib (http://quantlib.org/index.shtml). I am strongly leaning towards TA-Lib as I dont know of anyone at an individual level that would use all of the power of Quant Lib. That seems to be geared more towards high end quants & institutions.
I will draw up a process flow with what I envision as the whole financial ‘modeling’ process and then we can plug in the pieces with what is needed. I could envsion a 1)Data Feed Handler, 2) Technical Analysis Pre-Processors, 3) Messaging Queue (to initiate a buy or sell order from an execution platform) and probably some more operators that I cant think of at the moment.
Regards,
Eric
October 15th, 2008 at 7:01 pm
Tom & Eric,
It looks to me like TA-Lib is a Technical Analysis library (read indicators like MACD, RSI, Stochastics, etc) and Quantlib is about Quantitative finance and include a sha-load of very cool tools…
If your AI isn’t do’in TA internally then what is it doing in there? ;-)
Rapidminer + Quantlib = very interesting…
Throw in some nice data wrangling and your king of the world 8^)
Maybe what Rapidminer needs is a slick scripting language so you could hack up a io/pre/post processor without having to whip-out the entire devo build chain…
Cordially,
-Digital Dude-
“Good math and small teams win. -Alan Kay-
October 19th, 2008 at 8:06 am
DD: I’m not a comp sci guy, is a scripting language better?