Advanced Deep Reinforcement Learning System for Trade Execution: Part I: Foundation Concepts
Photo by Steve Johnson on Unsplash.com
As AI is taking over the world, it sure doesn’t stop when it comes to the realm of algo trading. Deep Reinforcement Learning (DRL) is already being employed in finance in a number of applications like trading bots, broker chat bots, risk optimization, price setting, or portfolio management.
Since DRL is a vast, complex subject I decided to create a learning series that examines the topic starting with the basic concepts with the end goal of building a full-featured DRL stock trading bot.
Ready to take a quick trip from the beginnings of Artificial Neural Networks to advanced Deep Learning Strategies?
This story is solely for general information purposes, and should not be relied upon for trading recommendations or financial advice. Source code and information is provided for educational purposes only, and should not be relied upon to make an investment decision. Please review my full cautionary guidance before continuing.
Artificial Neural Networks
Artificial Neural Networks (ANNs) were inspired by the biological neural networks of the human brain. They consist of interconnected nodes or neurons, which process information in a layered structure. Each neuron receives input, applies a function to it, and passes the output to the next layer.
This structure enables ANNs to learn and model complex patterns and relationships in data, making them ideal for tasks involving prediction, classification, and even decision-making in dynamic environments like financial markets.
Let’s take a look at a simple ANN architecture:
Input Layer: The Input Layer consists of neurons that receive numerical input data (also called features) and that have some predictive qualities in regard to the target value. In the context of algo trading, features may be OHLC prices, lags, deltas, technical indicators or news or social media sentiment.
Hidden Layer: The Hidden Layer consists of other sets of neurons that are interconnected with the input layer, receive data and perform computations like applying weights, using an activation function, or performing regularization tasks.
Output Layer: The Output Layer’s neurons are connected with the Hidden Layer neurons. It receives the processed data from the Hidden Layer and produces an output like a result or prediction.