Machine learning in trading: theory, models, practice and algo-trading - page 1935
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
1. then I still do not understand
3. there cannot be correlated signs after transformation. Yes, we can say that they merge into other structures (signs), but without redundancy
4.
it is all one expression, it is impossible to comment within it ))))
1. What is not clear here - we have our predictors, which we know how to reproduce, and then they are transformed according to certain rules and new predictors are produced which are calculated according to certain rules that emerged during transformation and which can be applied to any initial sample. So, I need to save these new rules to a separate file to convert a string when a new bar appears in the terminal, i.e. I need autonomy from R.
I'm saying that they were highly correlated before the conversion, which resulted in a significant reduction. But whether they merge or not - I need to analyze.
4. My R doesn't swear and everything works, but in essence of the question - model building and clustering are not related.
Here I downloaded an umap component, compressed 2500 attributes in one row, with a loss of course, but still for the tests it is enough
Think about it. I described the algorithm for 2500 features with a loss of +-10%.
So tell me how to unload these components :)
1. What is not clear here - we have our predictors, which we know how to reproduce, then we convert them according to some rules and get new predictors, which are calculated according to specific rules that emerged during conversion and which can be applied to any initial sample. So, I need to save these new rules into a separate file to convert a string when a new bar appears in the terminal, i.e. I need autonomy from R.
I'm saying that they were highly correlated before the conversion, which resulted in a significant reduction. But whether they merge or not - I need to analyze.
4. My R doesn't swear and everything works, but in essence of the question - model building and clustering are not related.
1. Well, then speak more clearly, because rules, rules are transformations, not rules, rules are if then else ...
you pass a trained model through the predict function and get the model output
4. not related, so what should I be looking at? what's the problem?
So tell me how to unload these components :)
PR <- predict(um , X)
PR <- predict( model , new data or whatever)
1. well, then speak more clearly, because rules, rules are transformations, not rules, rules are if then else ...
you pass a trained model through the predict function and you get the model output
4. not related, so what should I be looking at then? what's the problem?
1. Well how else to be clear - there are synonyms predictors/features/rules, we had one rule and converted to another, and I need conversion logic with which to get rules2 from rule1 - is that clear? :)
4. The wiki does say that:"dimensionality reduction is usually donebefore applyingk-nearest neighbor method ", and here I understand that clustering should be applied on new predictors/rules just to generalize obtained dimensions, so that they wouldn't be too many...
PR <- predict(um , X)
PR <- predict( model , new data or whatever)
And in what format should this be saved to a file?
1. Well, how else can I make it clearer - there are synonyms for predictors/features/rules, we had one rule and transformed into another, and I need the logic of transformation with the help of which you can get rules2 from rule1 - is that clear? :)
For those in the tank!!!
Synonyms are signs/features/predictors ...
Anything can be a sign, including a rule, but anything cannot be a rule.
Price may be a sign, but price is not a rule! WHAT IS NOT IMPORTANT ????
A function may be a sign but a function is not a rule; it is possible to construct particular rules based on a function but that is NOT the same as !!!!.
A mathematical transformation can be a sign of 2+2, but 2+2 is not a rule, it is a mathematical transformation. WHAT IS NOT IMPORTANT ????
And when you call a mathematical transformation a rule, a function you call a rule and are surprised that I do not understand you, then I get nervous, and I also get nervous when I read your post three times and try to understand what you mean. And when I tell you to express yourself clearly, and you start to argue, I begin to wonder why I need it all.
4. It says in the wiki that:"dimensionality reduction is usually donebefore applyingk-nearest neighbor method ", and here I understand that clustering should be applied on new predictors/rules just to generalize obtained dimensions, so that they wouldn't be too many...
read the curse of dimensionality, specially highlighted in red and in the largest letters .
And in what format do you save it to the file?
What format do you need?
This is from Expert, neuro bitcoin forecast for the last week, the first screenshot without filter, the second is the same, but with a "confidence" filter 40.
Euro is not of this quality, but close to it.
For those in the tank!!!
synonyms are signs/features/predictors...
Anything can be a sign, including a rule, but anything cannot be a rule.
Price may be a sign, but price is not a rule! WHAT IS NOT IMPORTANT ????
A function may be a sign but a function is not a rule; it is possible to construct particular rules based on a function but that is NOT the same as !!!!.
A mathematical transformation can be a sign of 2+2, but 2+2 is not a rule, it is a mathematical transformation. WHAT IS NOT IMPORTANT ????
And when you call a mathematical transformation a rule, a function you call a rule and are surprised that I do not understand you, then I get nervous, and I also get nervous when I read your post three times and try to understand what you mean. And when I tell you to express yourself more clearly, and you start arguing, I start to wonder if I need it at all...
Read the curse of dimensionality, I highlighted it in red and in capital letters.
Which one do you want?
All this rhetoric of individual perception.
Save it in a readable format, such as csv.
This python is an interesting thing. I have to make a stirring. I'm writing:
tmpIn=In tmpOut=Out I=np.arange(len(Out)) np.random.shuffle(I) for i in range(len(Out)): In[i]=tmpIn[I[i]] Out[i]=tmpOut[I[i]]Gives out bullshit. I reset it before looping In and Out and it works. I want to check every step.
1. It's all the rhetoric of individual perception.
2. Save it in a readable format, such as csv.
1. Fuck Alexey, if every teacher in your town had an individual perception, one would have the letter "B" as a "zyu". and the other one had a 69, you realize you still wouldn't be able to read!!! I read what you say, I do not understand, I get nervous, you do not get answers to your questions because I do not understand them, time is wasted, no use and who got better from this idiotic individuality of perception????
And what's holding you back?
Interesting thing about this python. Gotta do the mixing. Writing:
Gives out bullshit. I reset it before the In and Out cycle. It's working. You can check every step of the way.
What do you need this python for anyway?