Machine learning in trading: theory, models, practice and algo-trading - page 16
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
For financial BP we need to distinguish a pattern - that is, the monotonous behavior of BP over the entire available time interval.
Guys, I am attaching my advisor. It shows the logic of sending and receiving information from the R system.
To make it work you need to copy library from here: https://www.mql5.com/en/code/11112
There are instructions in file mt4R.mqh.
Silly assertion. EURUSD a year ago and EURUSD now are two different BPs with the same name.
Profanity. If your EAs don't hold up on the EURUSD, they are losing a year out of sample, it means you haven't found a pattern. You are trading noise.
Of course, your student references to the comparison of apples with oranges are valid for any other tasks, but you clearly can't distinguish these fruits on the Market.You want to find correlations in one single VR. And you want to find such correlations that must be present at all times in that VR.
These two circumstances (in bold) seem strange, to say the least.
Exactly as highlighted in bold. For you it seems strange, while several people on this branch, are trying to solve these very problems. Moreover, we are not the only people in the world on this branch, as a lot of knowledge-intensive tools have been invented and implemented before us, which allow us to solve these very problems.
So.
1.be present at all times.
This is the problem of retraining (overfitting) the model. If our model was able to identify some patterns on the available historical data, which are guaranteed to occur in the future, then we obtain yourpresence at any time. This can be achieved by building the model on predictors that are related to the target variable. Here is a tool called "main components" (the tool is quite old) which allows to sift out garbage (noise) from predictors that in the future will have the same patterns that we find in the available historical data.
2. You want to find correlations in onesingle BP
There is a fairly large set of different tools with different ideas behind them to find correlations. Here we are discussing NS, more specifically nnet. In my experience it is the least effective algorithm. Much more effective, and most importantly more visual, ada, randonforest, SVM - in descending order of performance.
Let's take randonforest as the most obvious.
What is the idea?
For example, having values of predictors we teach the algorithm to predict BUY and SELL. The algorithm builds a tree - combination of predictor values, belonging to one bar. One tree predicts BUY, the other one SELL. If we input about 5000 bars, the algorithm will find 200-300 tree variations. Further increasing the number of bars does not increase the number of trees. If we solve the problem on point 1, the resulting model will predict in the future with about the same error as on historical data.
For the theorists with experience it remains only to wish good luck in practice.
Thank you, darling.
We have several years of practice, too. That's why we're turning to theory. You just haven't grasped the grain of truth yet. Good luck to you too.
The first lessons look more like tutorials on this very framework than on data analysis. However, the leaders looked adequate, without the typical "I'm a Forex guru, I'll open your eyes and you will make millions" like in many other useless trainings, which gives hope that they will be telling adequate things till the end.
This is udacity, there will definitely be no bullshit there.
Pandas is sort of one of the most popular libs for data mining, python itself is a very handy language for a wide range of tasks.
No one is going to teach profitable trading. They are going to teach you how to take data, build a model on it and evaluate the result of the model.
This is udacity, there will definitely be no bullshit there.
Pandas is sort of one of the most popular libs for data mining, and python itself is a very handy language for a wide range of tasks.
No one is going to teach profitable trading. They are going to teach how to take data, build a model on it and evaluate the result of the model.
If you are a theoretician with experience, you can only wish luck in practice.
My experience as a stock speculator began with Borovoy's checks. Before that, I spent another 20 years investing in the real sector.
And you were born with the checks?
Anton Zverev
Let's not talk that way, people who learn and share their experiences here are willing to help each other, while you take the position of saying you are stupid here and I know everything) You'd better help me understand what you think and experience is right.
I agree with you that just giving BP is not enough, you need to compress the information and discard unnecessary things that prevent making the right decision, ideally to 0 or 1 those buy/sell, that isIf we have 10 indicators (that I do not believe in) and we have cut off 9 of them, for example RSI, we will still have too little, because Indus has its range and it turns out it does not work with values from -70 to 70, so it needs to compress and so on.. the question is how to do it?
I have thoughts about it, but I don't have enough knowledge to implement such a selector, yet...
My first try was a long time ago, I looped back from the current price and searched for a nearly identical situation in the past, then these situations were sorted by the result, how they ended up, for example, I had a current situation and found 10 analogues for it in the past 8 analogues ended with a price increase, 2 ended with a fall, so it will grow ... But the horror )) is that it was vice versa, the price often and strongly fell with these situations, with a strong bias in the buy direction, and then often retested tick by tick...
Then I created this kind of indicator, I took the cumulative sum of all buy prices and also the sum for profit, made their difference and got a certain index, when I compared it to the price it turned out to be opposite to the price, the correlation was -0.7 to -0.9 , so simply speaking the market goes against its own statistics, this is something to think about and reconsider