Discussing the article: "Neural Networks in Trading: Recurrent Modeling of Market Micromovements (Conclusion)" - 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
......
2. Study expert. I selected the same time period as specified in the expert. (Is it possible or necessary to change this, and should the period in the expert match the one I have selected, which differs from yours?)
....
When training models, you can experiment with the timeframe to achieve the best possible result. The expert advisor uses the timeframe specified in its parameters as a reference, rather than being tied to the timeframe of the chart on which it is running. During testing, it is advisable to specify the timeframe on which the model was trained.
......
Now, regarding this project. I’ll describe the stages of the launch, and please do correct me if I’m wrong:
1. EURUSD H1
2. The ‘Study’ expert advisor. I selected the same time frame as specified in the expert advisor. (Is it possible or necessary to change this, and should the time frame in the expert advisor match the one I’ve chosen, which differs from yours?)
3. Four files with the extension *.nnw have been created in the ‘files’ folder
4. I ran StudyOnline over the same time period (if I’ve understood correctly). Trades appear. Finished.
5. I ran Test on the test period. And here there’s just a solid line representing the balance – not a single trade.
What am I doing wrong?
I used the same time frames as specified in the article to achieve your result.
At this stage, I leave only one agent active in the tester and run a slow optimisation based on the agent index (Agent). This parameter is not used in the programme and is included solely for training the model in the strategy tester. I set the number of iterations to 200–300 and monitor how the results evolve. We only proceed to model testing once several consecutive iterations have produced similar results.
When training models, you can experiment with the timeframe to achieve the best possible result. The expert advisor uses the timeframe specified in the parameters, rather than being tied to the timeframe of the chart on which it is running. During backtesting, it is advisable to specify the timeframe on which the model was trained.
Good afternoon,
First of all, thank you for your interest in my work.
The ‘Experts’ folder allows for subfolders, which minimises confusion even when experts have identical names. For example, this is what it looks like in my terminal
What about the timeframe specified in the expert advisor? Do I need to change it to match the one I choose?
The training process is iterative and time-consuming. During initial training, I run Study dozens of times. The total number of epochs (the ‘Epochs’ parameter) far exceeds 1,000. By default, the advisor is set to 100 to save interim results and allow visual monitoring of the training process. With each run, the final errors logged should decrease. When they stabilise, we move on to StudyOnline.
At this stage, I leave only one agent active in the tester and run a slow optimisation based on the agent index (Agent). This parameter is not used in the programme and is included solely for training the model in the strategy tester. I set the number of iterations to 200–300 and monitor how the results change over time. We only move on to testing the model once several consecutive iterations have produced similar results.
Where can I check the training errors? In the log, there’s ‘train-->277’ – is that it?
And there’s also this in the tester’s log: ‘removed itself on 0% of the testing interval’ – is that normal?
Where can I check the training errors? In the log, there’s ‘train-->277’ – is that it?
And there’s also this in the tester’s log: ‘removed itself on 0% of the testing interval’ – is that normal?
Messages
Posts