Neural Trader (AI Project)
Published on April 10, 2025
Neural Trader is a neural network–based financial advisor developed after earning my Machine Learning certification from Matlab. The project uses Matlab’s Neural Network Toolbox to analyze current stock data.
A custom Python script scrapes daily data from Yahoo Finance and Google Trends into a CSV file, which is then processed in Matlab through a bagged trees algorithm. This method proved to be the most accurate, outperforming both convolutional and sequential neural network approaches.
In a backtest from January 2018 to December 2022, Neural Trader showed an average monthly portfolio increase of about 1%, significantly outperforming the S&P500 over the same period.
