Our technical approach is organized into four core components, shown below:
We began by aggregating data from a wide range of Data Sources to train and evaluate our machine learning models. The Project Architecture provides an end-to-end blueprint of our system design, showing how the components work together to bring our vision to life. The Sentiment Model uses news articles to forecast shifts in market sentiment and inform market timing decisions. Finally, the Deep Learning Model integrates sentiment insights with additional security features to generate optimized security selection across the user's stock universe.