[ad_1]
- Video Views: 7696
- Published On: 2021-02-08 21:55:17
- Video Published/Author: Python Lessons
- Video Duration: 00:30:31
- Source: Watch on YouTube
In this tutorial, we’ll learn what is the simplest way to download historical cryptocurrency OHCL market data via exchange APIs and we’ll use this data to train our Reinforcement Learning Bitcoin trading agent that could finally beat the market!
Text version tutorial: https://pylessons.com/RL-BTC-BOT-Historical-data/
GitHub code: https://github.com/pythonlessons/RL-Bitcoin-trading-bot/tree/main/RL-Bitcoin-trading-bot_6
✅ Support My Channel Through Patreon:
https://www.patreon.com/PyLessons
✅ One-Time Contribution Through PayPal:
https://www.paypal.com/paypalme/PyLessons
[ad_2]
I got lost at around 4:09. What is that app sir? Sorry I'm a beginner.
I got this working on Binance and Robinhood data. Thank you!!!
Can you do the same for binance futures?
this line doesn't work no more: api_v2 = bitfinex.bitfinex_v2.api_v2()
AttributeError: module 'bitfinex' has no attribute 'bitfinex_v2'
how should we solve it?
Thanks for uploading this video, it helped in clarify a lot of things for me. I have the following question: How do you combine the learning of these parallel agents? Do we take the weights from the agent that performed best? Which part of the code does this as I was not able to find where we are combining the weights? I think without combining, it is same as running a single worker.
can you please try it with binance
hi the video is awesome and it helped me a lot but when I want to download those cryptocurrencies which have less than 3 words or more in their symbols it won't download such as BUSD, USDt , DOGE,… what should I do please help me.
where does the data go when i download it, i dont want my
ssd to get spammed with
trainings data
Amazing job!
I get a broken pipe error when attempting to train the model
Nice job! Look into risk/reward management. This will improve your model. Consider using a stoploss (in terms of % for example) and one or more take profits. If first take profit reached, consider moving stoploss to entry. If you have a risk reward of 1:2, this means you risk 1 to earn a reward of 2. Mathematically speaking you need to have 33.3% accuracy. Good luck, in any case.
Please add an option or go over how this could be applied to another exchange api (ie coinbase or kraken). Thank you.
Really enjoying the walk through and seeing the process of this as it develops.
Do you have a video explaining the steps in your reinforcement learning part?
Nice Video!