Tensorflow-NeuroEvolution-Trading-Bot

Uses neuroevolution instead of backpropagation, allowing trading strategies to emerge through simulated natural selection, as emphasized in the project philosophy for an alternative to gradient-based optimization.

Evolves models based on fitness from trading performance, eliminating the need for historical labeled datasets, which is ideal for environments where such data is scarce or unreliable.

Creates and iteratively improves a population of neural networks, increasing the chance of discovering robust strategies through mutation and selection, as described in the generational improvement process.

Trades cryptocurrency based on technical indicators, making it suitable for markets where price action and chart patterns are primary decision drivers, as stated in the key features.