Hey Musa! Actually, the project is still ongoing. I’ve some ideas in mind and have backtested some strategies, which resulted really profitable, but the next and last step is missing: building the actual bot to place orders for me.

When it comes to accuracy in this kind of models, it’s not a metric I would try to optimize. Because a model with a 20% accuracy can be more profitable than another one with a 70% rate of success. So, my advice is to aim for profitability straightaway, and a low variation if possible, but forget about accuracy.

Thanks for your comment!