Discussion of article "Random Decision Forest in Reinforcement learning" - page 2

New comment
 
Sorry Maxim, I was distracted, I spelt your name wrong.
 
Forgot to add. The solution, in strategy games, is combinatorics (search for repeated combinations), probabilistic analysis, determination of the opponent's psychotype and creation of a behavioural model.
 
Vasily Belozerov:
Forgot to add. The solution, in strategic games, is combinatorics (search for repeated combinations), probabilistic analysis, determination of the opponent's psychotype and creation of a behavioural model.

Greetings, about psychotype - certainly not applicable to the market. Otherwise, RL is largely a legacy of game theory. For example, you can think of several RL agents as competing players, each trying to beat the others. Alternatively, make them mutually reinforcing.

Essentially, a payment matrix between the market and the agent is drawn up. And if the market's strategy does not change, the agent starts to beat it, when possible.
 
I was just suggesting that maybe the right machine should have something not right added to it, similar to human actions. For example, first move e2 - e4, then second move e4 - e2, third move again e2 - e4. Basically, something like that.
 
Vasily Belozerov:
I was just suggesting that maybe the right machine should have something not right added to it, similar to human actions. For example, first move e2 - e4, then second move e4 - e2, third move again e2 - e4. Basically, something like that.

The task is not to make trading be similar to human trading (and how do you know that, the bot has to pass the Turing test?) and what makes you think that a human is able to trade well or better than an Expert Advisor, it all depends on the strategy. The task is that the machine itself would look for some patterns and trade them.

 
Vasily Belozerov:
I was just suggesting that maybe the right machine should have something not right added to it, similar to human actions. For example, first move e2 - e4, then second move e4 - e2, third move again e2 - e4. You know, something like that.

This is Random forest, its main point: that it forms the opening book by itself.

 
Maxim, is there any effect of trying to generate new random trees? For example, the first time we generated one tree, the second time we generated other trees, which are more efficiently trained. Or does it have no effect at all on the final result of training?
 
Ivan Gurov:
Maxim, is there any effect of trying to generate new random trees? For example, the first time we generated one tree, the second time we generated other trees, which are more efficiently trained. Or does it not affect the final result of training at all?
It's not the trees, but rather training multiple forests on the same data yes, it does. Because the construction process is randomised and the results may vary. I was surprised that an ensemble of forests gives a noticeable improvement, i.e. train several forests on the same data and average the result. (5-15 pieces) You can make different settings too. Even better results should be in case of bousting, but I haven't got there yet.
 
I liked the article. I wish someone would redesign the forest in alglib so that you can adjust the depth of the trees.
 

Thank you for sharing this very useful article.

I was trying to implement additional indicators to the code, but I am not an expert programmer and not good experience as how to use membership functions and hence, I couldn't get my head as how to add more indicators to be used along with the rules inside the OnInit() function. The code contains only RSI indicator and creates the BUY and SELL rules out of that. Can you please provide few more example codes of indicators like Moving average or MACD or stochastic OR SAR to be used in the code?

Especially, I want to know as how to create rules and add to entry conditions while comparing it with current price. The main problem with current code is that sometimes it holds loosing trades for long time while closing profitable trades quickly and so any advice on this will be appreciated. I think more filtering on exit logic needs to be done.

Also, I have one question if you can answer please:

Does the OPT file continuously updates in order to improve the entries and exits over time after long time by fine tuning the policy itself?

Or does the EA just uses the strategy tester to optimise the EA values and uses the same entry and exit values of which were profitable recently like regular optimised EA?

I mean like other Neural network EAs, does it fine tunes it's overall policy for trade entry and exit during the course of trading?

12345678910
New comment