Gathering a team to develop an IO (decision tree/forest) in relation to trend strategies - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Thank you ! I'll be learning the trees.
You're welcome. There are a lot of theoretical lectures on the Internet, there are also examples in Python, it makes sense to watch different lecturers, because the topic is wide and everyone can add something clever.
Recently I saw this short presentation by a student(?), I liked it - brief and a lot of information.
Interesting,
I would join later, now I have to deal with mine, more ideas than hands and skills
Alexei, hire a designer for your ava
What task is unclear to you? Why are you criticising? Don't you remember that I put Rattle on, looked and felt it. I have concluded that the standard approach is no good and a unique method for trading is required. Strange that over the years you haven't realised this.
I did,
I'm off,
Good luck with reinventing bicycles.
Interesting,
I would join later, now I have to deal with mine, more ideas than hands and skills
Alexei, hire a designer for your ava.
I understand - everyone has his own cockroaches (ideas).
Thanks for the willingness to participate - it's at least a little reassuring. And then look there is no criticism on the merits, no desire to participate. Usually the creators of such threads are scolded for the lack of goals and plans, I kind of gave it all, but the effect is even less pronounced.
Detached. The funny thing is that everyone thinks they have a brilliant idea and instead of speeding up its development they hide it for years and stew in their own juices. I have long understood that I'm not a genius, I don't have a grail or a super idea, but I do have a desire to create a platform where people can pool their knowledge to achieve common goals on a particular issue.
However people like SanSanych Fomenko, who certainly has the knowledge needed for this project, will sooner say that my views on the problem are wrong than objectively evaluate them or offer alternative approaches.
One can gather only crumbs of knowledge from open sources, including solution algorithms. I'm not suggesting to reinvent something super new, but to adapt methods to the data being processed and to get an efficiency gain from it.
When there is a question of classification, two or more objects are estimated by characteristics, but in trading this approach is insufficient - since it is not enough to classify a trade result by profit or loss, there can exist a situation when 70% or more percent of sample is classified correctly, but we incur losses. I prefer to classify event probabilities on each opening of a new bar, which gives an acceptable sample for learning, but sometimes leads to paradoxes when a good classification gives an aggregate loss. That is why I propose to think about additional criteria for a fitness function, which would express not only the quality of classification, but also give the evaluation in quantitative terms of classification efficiency. And there are many such nuances, and it is these that I propose to address.
I remembered,
I'm off,
good luck inventing bicycles.
Thank you.
Well, why go away - buy some popcorn and laugh at the fool!
There are special MO methods for working, for example, with graphic images, there are also decision trees and NS, but there are additional ideas - specialization, this is the specialization I propose for trading, not proposing to re-invent the wheel.
The excitement around the grail continues unabated.
Are you referring to the graphic on your avatar? It's not clear what you mean - the MO is far from a grail, just a method of finding a solution.
Aleksey Vyazmikin:
Otherwise, I see no substantive criticism and no desire to participate. Usually the creators of such threads are scolded for lack of goals and plans, I kind of gave it all, but the effect is even less pronounced.
alas, your plans are of no interest to anyone - such a plan any 30+ person sketches in 10 minutes in three approaches )))).
To attract the public, besides plans we need results, any results, at least in trading, at least in a sound mathematical model, even implemented in code. In a neighboring thread about scalping - people showed up in half a day, and mind you, the author simply stated his goals and gave a working model.
but with plans like the one in the first post.... You should hire hired workers for pay - they will flap their ears and look in their eyes and wait for their paycheck))))
i don't want to offend the topicstarter, as it's fashionable these days.... that's just my judgement!!!
Alas, no one is interested in your plans - any 30+ person could sketch such a plan in 10 minutes in three attempts ))))
Are we talking about the proposed areas for working together?
If 30+ people have the same plans, what prevents them from teaming up? Or is it about something else?
If you want to attract the audience, besides plans, you need results, any results, either in trading, or in a well-founded mathematical model, or implemented in the form of a code.
I have shown intermediate results of learning by my method, even in the tester.
Well, with plans like the one in the first post.... they will clap their ears, make loyal eye contact and wait for their pay.)
Again, what are the plans being discussed - the direction of joint activities? And what is expected - the delivery of a ready-made solution out of a box that just needs frills?
I'm not squeamish about paying people hired there to solve problems. I just really understand that the project is not simple, and it is cost-effective to combine to move towards the goal of adapting the method to the specific tasks at hand.
I do not want to offend the topicstarter, as it is fashionable nowadays. it's just my value judgement!!!
Why should I be offended - well, except for unconstructive criticism...