February 2010 Archives

One of the video tutorials I'm working on is how to build and use a direct mail marketing model using a Decision Tree learner.  Rapidminer can be used for small, medium, and larger marketers as a data discovery tool that helps them understand the buying habits of their customers, and where to target their efforts.  A Decision Tree model helps segment your market and identify the "sweet spot" in your customer base.

Check back over the next two weeks for this video tutorial.  I'll show you how to mine your data and start being a more effective marketer in under 10 minutes.

In this video I discuss how to use a cross and simple validation operator to split your training data into two sets: training and validation data sets.  I also highlight the new intuitive “quick fix” error solution suggestions in Rapidminer 5.0. Enjoy!

Video download link (HQ): Rapidminer 5.0 Video Tutorial #3

See the Rapidminer 5.0 Video Tutorial #2 post for the data files used in this video

My readers may or may not know but I maintain a professional profile on LinkedIn (you can find me at Thomas Ott). I really like LinkedIn and I’m always looking to network with good professionals in the engineering, data mining/modeling, and financial world. Please feel free to reach out to me because I would love to connect with you! 

But I do ask for a small favor in return...

If you are a reader of Neural Market Trends, and you enjoy my video tutorials and content, please take a moment to write a recommendation for the site. Your candid thoughts and recommendations are very important to the success of this blog and it allows me to grow this site by providing my readers with free Rapidminer tutorials.

The information I share on this site via my videos and tutorials is extremely valuable and free, there’s nothing like this available on the Internet that I know of. Please consider my above request, it would be greatly appreciated.

Looks like I'm on a roll!  Please see my Rapidminer 5.0 Video Tutorial #2. In this video we begin the process of recreating my original written NMT YALE/Rapidminer tutorials into version 5.0 and into a video.  This video shows how to import training and prediction data, add a classification learner, and apply the model, and get the results.

Video download link (HQ): Rapidminer 5.0 Video Tutorial #2

The data files you will need to follow along are Excel spreadsheets below:

Training data set: gold_final_input

Prediction data set: ga-gold

I've been updating my social media links for a new and improved roll out of Neural Market Trends over the next few months.  My first job was to create a Youtube channel for my new Rapidminer videos, which I anticipate to update on a weekly basis until I've redone all my previous tutorials and added some new ones. I expect that amount to be about 10 or so videos.

Next I created a Facebook Fan Page for my readers to join.  Please note, I usually don't add thousands of friends on FB so if you request me as a friend and I don't add you, please don't be angry or upset.  I'm just very protective and secretive with my inner circle.  Please, instead, join the Facebook Fan Page and interact with me there.

I'm also trying out the Google Friend Connect, which you see in the right sidebar, and I'm available via GTalk.  Please note, that I'm usually very busy from 7AM to 5PM Eastern Time, so chatting is limited to non-existent that time. If you catch me before or after that time, ask me questions, chat with me about Rapidminer ideas, or just say hello.

And finally, I'm still on Twitter but only use it to broadcast my blog posts.  It's too noisy for me.

2010 is going to rock!

Long time linker (and reader hopefully) Max Dama maintains a great blog on automated trading and provides lots of great examples on how to do this.  I recently discovered that he has quite a few Python programming examples as well.  I'll have to spend more time there again digging through his posts and I suggest my readers do too.

$SLV seems to have tricked us after bouncing and closing above the 50 WMA last week.  That's good news and it appears SLV will be gaining steam this week as it rides up trend channel.  If $SLV closes above the 50 WMA this week, then I'd see we're in the clear until we're not.

Video download link: Rapidminer 5.0 Video Tutorial #1 PS: I'm glad to be back guys. Leave me a comment if you want more, please stroke my fragile ego. LOL. PPS: My Youtube Channel is here: Neuralmarkettrends1 PPPS: For those who want to follow along, see the original GE.xls file.

I downloaded and started to fool around with another genetic algorithm addin for Excel.  This one is called xlbit and is created by Xlpert.  What I like about it is that it has a better interface than Gentik Solver, but its not free. It allows you do max and min for target cells, add constraints (a big plus), and lets you choose between three genetic algorithms: Elitism, Roulette, and Tournament.  It even creates a new worksheet with results and graph for your average and overall fitness.  Sweet!

If you like it, and I do, you can buy the addin for $15.  That's pretty reasonable considering the developer packages this addin with two more addins, a neural net and a type of value at risk routine.   Xlbit does come in a demo version, so you can try it out first.  

So what has your intrepid neural net / genetic algo techno trader been doing with this addin?  Writing new trading models? Analyzing the market? Finding hidden trading opportunities? No, not at all.  I've been genetic modeling the perfect measurements of women.

Blow chunks, suck donkey dick, or  whatever metaphor you want to use if it breaks through and closes below $1.35.  Could it be temporary? I don't know but the daily chart (not shown in this post) already shows the 50 DMA below the 200 DMA and gaining traction.

The weekly chart has the 50 WMA awfully close to the 200 WMA and getting closer.

In case you didn't know, Rapidminer 5.0 has been released and there have been some MAJOR changes and improvements to it.  It looks a lot easier and appears to be even more powerful.  I just downloaded it and plan on playing around with it for a while.

Update: Rapidminer 5.0 is beautiful, just simply beautiful!  The have a new GUI that is to die for and a simple drag and drop for your data imports.  With all these changes, all my tutorials will need to be completely updated... (hint, hint)

$DNDN is routinely making all time highs and the question I pose this morning is, would you buy it now?

 A lot of people are scared to buy all time highs. They think that since its at an all time high, it can't go higher and maybe its run its course.  

From the weekly chart you can see that it just keeps on making all time highs. 

But the reality is that ONE day it will have reached its all time high and then sell off. The real question then becomes, when has it truly reached its all time high?  No one knows the answer to that question!

So the moral of THIS story is to always use a stop IF you buy an all time high, because you never know if THAT all time high was ITS last!

Well all the news about the PIIGS taking wing in Europe have rattled the EURUSD pair.  I think it might be heading down to test the $1.25 level and I'll be keeping a close eye on it, since I'm about to start trading Forex in earnest again.

I have my models all rebuilt and my system ready to go, now its just a matter of moving to a new broker and setting up Metatrader 4.  I'm going to start fooling around with Metatrader as well as testing some expert advisors because rules based trading appeals to me and my neural nets.

I'll be starting a NEW $100 Forex Experiment and seeing how quickly I can build it up or blow out!

In January of this year, I decided to ramp up my posting frequency again for Neural Market Trends.  Part of the reason is to data mine blog traffic data for the express purpose of learning more about SEO and Internet Marketing.  I started this doing this after I had a few beers with the Market Doctor and we discussed the long tail statistics of Internet Marketing and how some blogs just make money hand over fist.

So I'm going to do a little experiment and see if I can use neural nets to maximize SEO and Internet Marketing.  January 2010 will serve as my starting metric, let's see if I can break $1000 a month by December 2010.

Channel Report:

• Google Adsense: $11.17
• Text Link Ads:        $29.42
• Adify:                        $3.65
• Affiliates:                 $0.00
• Total:                      $44.24

The blog had 2,772 visits and 6,744 page views in January.  Average time on site was 4:29 minutes.

$SLV is acting strangely again. A warning flag came up in my model that the trend might be breaking down in this ETF BUT it sure looked that way back in July 2009 as well.  Then it broke below the channel but stopped at the 50 week moving average. This time it poked and closed below the 50 week moving average.

Is $SLV tricking us again?  What do you think? If you leave a comment, I'll reveal my answer...

Well would you? Or would you wait till $1021ish?  What's everyone's long term outlook on Gold?