Neural Networks

Quote from HoundDogOne:

...
(If you are a computer genius and skip steps #1 and #2...
You will likely have a very sub-optimal black box system...
And will be competing against algo systems built by professional trading firms...
Firms with multi-million dollar tech budgets...
And a zillion man-years of hardcore trading experience.
More...

rcanfiel and HoundDogOne,

Do you have any clue about what you are talking about? Who are you to tell what is possible and what is not? It is not becasue you weren't able to do it that it is not possible...
I reiterate that "I DO NOT know how to trade manually" but I know a bunch about statisitcs and programming. Check some 1 or 2 years old posts I did about ATS performance.
Who cares if my blackbox is optimal or sub-optimal. All I care is that I have been using it for over 4 years and that my programme have been making more money trading than I was making working at a Fortune500 technology company.
The software is allowing me to quit my job easily and pay for a top-tier university to do another masters in Machine learning/Financial engineering.
So maybe you are right, but it is not relevant as long as my programme makes me enough money to have a comfortable living.
Cheers,
 
To add to that, one only has to view the results of the MQL trading competition to see that AI has real value in trading, its all there in black & white.
 
Quote from Craig66:

To add to that, one only has to view the results of the MQL trading competition to see that AI has real value in trading, its all there in black & white.
More...

Yes, the median participant of the 600+ players has traded his account from $10,000 into ~$5200 in only 8 weeks!! 119 players have actually increased their accounts!!
 
My regular work involves machine learning and data mining.

At home I use a combination of NN's, GA's and Tree classifiers.

To be honest, my initial analysis of any data almost always involves a Tree classifier (depends on the data).

Look into WEKA (open source Java machine learning framework).

Like someone else posted, these are tools. The deciding factor is you. You have many decisions to make, i.e. time slice, data pre-processing, etc etc.
 
Quote from markettrak:

rcanfiel: My log shows that you visited my site a total of 15 times during two one-week trial subscriptions. You were never a paid subscriber. The hardly qualifies as a "long time."

My model returned over 20 percent this year while the market is up about 5.

Rich
More...

You must be joking. I visited your site many time, You used to post signals for free for a long time a while back.

Notice that you don't provide a serious track record or archive of all your past calls??

Your system does not work, and should be avoided.
 
Quote from StocksSniper:

rcanfiel and HoundDogOne,

Do you have any clue about what you are talking about? Who are you to tell what is possible and what is not?

Since you point to yourself as a "successful example" it is quite apparent you know little. I minored in AI (machine vision, robotics, plannnig/understnading, LISP, expert systems, etc.) as part of an MS in computer science. Your credentials??

Who are you to question people you don't know, assuming you know their history with AI & Neural It is not becasue you weren't able to do it that it is not possible...

I did a lot of research on neural nets, on studies involving them, forums, and other places for over 5 years. The general consensus was that they were generally worthless, as is your own testimony, as you seem to not understand the native dificulties with trying to make one serious and profitable, long-term.

I reiterate that "I DO NOT know how to trade manually" but I know a bunch about statisitcs and programming. Check some 1 or 2 years old posts I did about ATS performance.
Who cares if my blackbox is optimal or sub-optimal. All I care is that I have been using it for over 4 years and that my programme have been making more money trading than I was making working at a Fortune500 technology company.

Anecdotal testimony from an unknown person without any trackable evidence is valueless. If you really understood statistics, I shouldn't have to tell you that.

The software is allowing me to quit my job easily and pay for a top-tier university to do another masters in Machine learning/Financial engineering.

This does not impress, if you are trying to win brownie points.

[/B]
More...
 
Quote from rcanfiel:

More...

I am not trying to impress you or anybody else as I do not know you and I have nothing to gain. You have a MS in Computer Science? So what? I have one as well... Maybe you need to go back to school and relearn your stuff.

Is it better to trade manually? I do not know... But I know that I COULDN'T do it. Does that mean nobody can trade manually? Of course not. Maybe others can, but I can't.

the same goes for you and your pals. You tried to use AI, but you failed and you concluded it is useless. You need to come down to earth and realize that other people can use it differently and succeed.

Your incompetence in AI doesn't prove anything besides that you do not know how to use it properly. My competence proves that you can use AI only and make money.
My incompetence in trading manually doesn't prove anything besides that I do not know how to do it. Your "success" (IF you are successful -this is A BIG IF) proves that it can be done.

Cheers,
 
Quote from StocksSniper:

I am not trying to impress you or anybody else as I do not know you and I have nothing to gain. You have a MS in Computer Science? So what? I have one as well... Maybe you need to go back to school and relearn your stuff.

Is it better to trade manually? I do not know... But I know that I COULDN'T do it. Does that mean nobody can trade manually? Of course not. Maybe others can, but I can't.

the same goes for you and your pals. You tried to use AI, but you failed and you concluded it is useless. You need to come down to earth and realize that other people can use it differently and succeed.

Your incompetence in AI doesn't prove anything besides that you do not know how to use it properly. My competence proves that you can use AI only and make money.
My incompetence in trading manually doesn't prove anything besides that I do not know how to do it. Your "success" (IF you are successful -this is A BIG IF) proves that it can be done.

Cheers,
More...

I have a degree in Computer Science (1982)...
And 15 years as a full-time trader...
Manage a 7 figure hedge fund...
And do about 3,500,000 shares/month.

So I have DUAL professional expertise...
And might have a different perspective then someone with more narrow experience.

There is no hard information from the AI crowd in this thread...
It's just a bunch of claims...
And the website mentioned (markrttrak)... is a joke.

Generally...
The field of AI involves modeling the most advanced intelligent entity there is...
The HUMAN BRAIN.

The only way to accomplish this in trading...
Is to model the thinking process of a human Expert Trader or Market Maker.

Something like a CBR approach to build an expert system...
Makes sense...
Especially since well-designed CBR systems have produced near pro level Poker Bots.

But, IMO, using sexy neural nets or immunologic "learning systems"...
WITHOUT ANY ACTUAL ACCESS TO A HUMAN EXPERT...
Will produce very marginal results.

This is akin to building Medical Diagnosis Expert Systems...
WITHOUT ANY INPUT FROM A MEDICAL DOCTOR.

You are better off in the long run...
Spending a few years becoming an Expert Trader...
Than just dogmatically building sub-optimal AI systems...
Which will be competing against Human Experts running ** far more optimal ** systems.
 
Quote from HoundDogOne:
I have a degree in Computer Science (1982)...
And 15 years as a full-time trader...
Manage a 7 figure hedge fund...
And do about 3,500,000 shares/month.
More...

Good for you and I am happy for you. Not trying to be sarcastic or anything. I am really happy for you. Managing a hedge fund must be difficult and I give you credit you doing it.
I have a LOW 7 figures account and it is fully managed by my software.

Quote from HoundDogOne:

Generally...
The field of AI involves modeling the most advanced intelligent entity there is...
The HUMAN BRAIN.

The only way to accomplish this in trading...
Is to model the thinking process of a human Expert Trader or Market Maker.
More...

Wrong. Very wrong. This is very far from being "the only way". Maybe that is how they used to teach it in 1982, but today "System Experts" are just a portion of the AI field. Most of AI nowadays is based on statistical models.
Check the book I mentioned earlier in this thread "The elements of statistical learning by Hastie" as an example. It is all about AI, but doesn't talk once about System Experts.
The MS in AI at Stanford have few classes in System Experts and many more in Statistics (Check their wbsite).

Quote from HoundDogOne:
This is akin to building Medical Diagnosis Expert Systems...
WITHOUT ANY INPUT FROM A MEDICAL DOCTOR.
More...
Wrong again. What we are talking about is building a "search engine" and classifying data without understanding the rules that generate that data (Google? Rings any bells?...)
 
Quote from HoundDogOne:


But, IMO, using sexy neural nets or immunologic "learning systems"...
WITHOUT ANY ACTUAL ACCESS TO A HUMAN EXPERT...
Will produce very marginal results.
More...
valid point
 
You must log in or register to reply here.
Back
Top