Trading robot using neural networks - Finam Trade API Hackathon
A trading robot that uses computer vision-based neural networks to identify specific patterns on stock charts and executes trades using the best trained model.
This example of a trading robot using neural networks is well documented and goes through all the steps step by step:
obtaining historical stock data
preparing a dataset with images of formations from a stock chart using a specific logic
training the neural network and selecting the best trained model using the parameters loss, accuracy, val_loss, and val_accuracy
verifying the predictions made by the neural network
verifying the connection to the Finam API
launching a live strategy using the selected best model of the trained neural network
Author: Oleg Shpagin
Github: https://github.com/WISEPLAT
Telegram: https://t.me/OlegSh777
Link to the code:
https://github.com/WISEPLAT/Hackathon...
Terms of Use
The program code, available online at https://github.com/WISEPLAT/Hackathon..., allows for trading operations on The Stock Market Using Neural Networks program is created solely for the convenience of working with and studying the applications of neural networks/artificial intelligence. When using the Program, the User must comply with the provisions of the current legislation of the Russian Federation or their country. Use of the Program is offered on an "AS IS" basis. No warranties, either oral or written, are included or implied. The author and the community do not guarantee that all errors in the Program have been corrected. Accordingly, the author and the community assume no liability for the consequences of using the Program, including, but not limited to, any damage to the User's equipment, computers, mobile devices, or software caused by or related to the use of the Program, as well as for any financial losses incurred by the User as a result of using the Program. No one is liable for data loss, damages, losses, including incidental or consequential, lost profits, lost revenue, or any other losses related to the use of the Program.
The Program is distributed under the terms of the MIT License.