search

Netanelshoshan / freqAI-LSTM

A Trading Model Utilizing a Dynamic Weighting and Aggregate Scoring System with LSTM Networks
MIT License
82 stars 25 forks source link

Regarding chart patterns and other possible inputs #10

Closed ckaraca closed 1 month ago

ckaraca commented 1 month ago

Thank you for sharing the FreqAI implementation. As an enhancement, is it possible to populate chart patterns and train the AI with the additional inputs? Or how can we feed it with real-life data such as SPX and DXY?

The image represents an indicator on Tradingview drawing the Cup and Handle formation. image

Netanelshoshan commented 1 month ago

Hi @ckaraca,

It's something I would really want to see in the futute and depends on CCXT / exchange support.

This repo was initially a Tensorflow plugin for Freqtrade then I ported it to PyTorch and was served as a "how to". The example strategy was something I uploaded bc asked to give an example. You can add / rewrite your implementation in populate_indicators / set_freqai_targets() and add that to final target "T"

I highly recommend you rewrite the strategy, fine-tune the weights with one of the loss-functions that Freqtrade offer or implement a custom one yourself (they use scikit-learn in the backend).

As much as all of us wants to find the perfect strategy - ALWAYS test your strategy before you push it to prod (aka live).