Gathering a team to develop an IO (decision tree/forest) in relation to trend strategies - page 11
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
Aleksey Vyazmikin:
What prevented me personally from understanding something was the lack of a help in Russian - I saw words I didn't know and closed everything down.
Yeah... With such a desire for new knowledge this caravan is going nowhere...
I'm stocking up on popcorn.
To achieve the goal from the first post you need to:
But it seems the author of this thread has a completely different goal...
I (personally) don't have a goal to learn new programming languages or technologies (to customize them for consumption), and I haven't reported that here.
Again, the goal is to create a team of people who either or individually:
1. have fresh ideas
2. Have knowledge of programming languages
3. Willing to provide any resources, or just invest
4. have knowledge of MoD.
5. Have knowledge in other areas but are willing to apply/convert their knowledge to a common goal - maths, statistics or whatever
6. Others that are willing to be useful to the project, such as collecting and analyzing data.
7. I think it would be good to make a separate portal for the project, so a good professional in this case would also be useful and be able to use the achievements of the team.
Yeah... With such a desire for new knowledge, this caravan isn't going anywhere...
I am stocking up on popcorn.
You're a strange man. In fact, you probably did something clever, but no one appreciated it. I tried to tell you what the reason is but you don't want to hear it. I repeat that without instruction in Russian (the language for the target audience), with pictures and explanations your geniuses will not be assessed. I suggested you to show your talent in the group, where there will be people who are not only able to create something, but to explain and describe, for example Maxim. But you probably don't even like the idea that someone doing the same thing as you may get a different result.
So, the offer is open, I propose not to oppose ourselves to others, and to seek opportunities for harmonious coexistence in order to achieve a common goal.
I (personally) have no goal of learning new programming languages or technologies (setting them up for consumption), and I have not communicated that here.
Again, the goal is to create a team of people who either or separately:
So what exactly will you be doing? What is the use of your knowledge for the team?
Right now there is only a project manager who wants results without understanding the topic...
You're a strange man. What in fact - did something clever, probably, but no one appreciated it.
"I'm a programmer!"
Open Sources is for sharing experience, not for profit. At this stage of development of my library I need experienced programmers who are able and willing to understand the code. There are no such people yet...
So what exactly will you be doing? What is the benefit of your knowledge for the team?
Now there is only a project manager, who wants to get results without understanding the subject...
My benefit is in organising people, at the very least. On the first page, I have indicated what I can give.
Are not you surprised that according to your own statements you are such a clever and useful person, and I am such a stupid and useless person, suggested to do what you have already implemented in 90%? Or you are not so clever that even I, such a stupid person, could think of something you have implemented and decided that it is necessary, or ...
Once again, I spend many hours every day studying exactly on the issue of MoD, watching lectures, writing out ideas that can be applied, self-educating as best I can.
I don't care about my self in this project, what really matters to me is the achievement of a group of people who will all be able to benefit from the shared achievements together.
"I'm a programmer!"
Open Sources is for sharing experience, not for profit. At this stage of my library development I need experienced programmers who are able and willing to understand the code. So far there are no such people...
It's great that you're a programmer! I program only for myself, and I don't proudly hang up that title.
Regarding "they want to understand the code", as someone who actively seeks the help of intelligent people for a fee, I tell you that people don't want to understand other people's code even for money! It's generally believed that if they need to, they will write something of their own. Your task is not only to put out the code but to show that they need it - to attract such people. Do you feel that the vector is similar between us? This is why I propose to jointly develop trading-oriented MoD for all those, who are interested in it.
You still haven't googled Spark?
Do you read at all what is written to you? Or were you praying so hard for Spark last night that the exalted state has not yet passed. I say the forest is an obsolete model already, and you convert it back to alglib format, what the hell would be the point. Wouldn't it be easier to do everything in python? And you don't need special speeds there, if data is not properly prepared and the model is weak then no enumeration through genetics on hundreds of cores will not give anything.
And what's the "rue forest"? Is it classic? And when did it become obsolete?
Plugging in the DLL doesn't allow you to use the agent cloud and makes testing more difficult. Why give up the cloud when all implementations of the random forest algorithm can be losslessly converted to the simple and very fast AlgLib format?
Speed is always important. If you don't understand this, the scale of the research is small.
And what's the "rue forest"? Is it classic? And when did it become obsolete?
Plugging in the DLL doesn't allow you to use the agent cloud and makes testing more difficult. Why give up the cloud when all implementations of the random forest algorithm can be losslessly converted to the simple and very fast AlgLib format?
Speed is always important. If you don't understand this, the scope of the study is small.
To assess the quality of the model (network or forest), error values of MSE, OOB, etc. are used.
But unlike in picture recognition, where a human is the verification sample, it is very difficult to make such a sample for a price chart. Therefore the estimation of a price model by MSE, OOB, etc. is often misinterpreted.
There is no clear definition of "overfitting" the model.
So I stopped checking the model by MSE, OOB, etc.
I now prefer to overlay the training result on the price chart to see the whole picture.
Here is my method for evaluating the quality of the model (I have already posted this picture):