AI Investing with Brain

Published: April 26, 2023

C8 Weekly Bulletin:  AI Investing with Brain

26 April 2023


The C8 Weekly Bulletin returns after an Easter break with the announcement that Brain, an AI-driven research provider, has now joined the C8 platform. Brain has developed NODES; an integrated Artificial Intelligence Engine that combines Artificial Intelligence algorithms with proprietary Natural Language Processing techniques to process a large volume of external information.

Using NODES, Brain build alternative datasets, asset allocation signals and investment strategies, as well as Artificial Intelligence based products and services for the Financial Industry such as thematic selection or smart event-detection systems. Please find below two examples of the indices that can now be found on C8 Studio, do contact C8 if you would like to know more:

Brain Sector Rotation Based on News Sentiment

  • The index selects the top three US equity sectors (using the corresponding US ETFs as proxies) with the highest sentiment from financial news in the previous few months. The selection process is based on an Artificial Intelligence proprietary engine that scans financial news from thousands of financial media sources.
  • The system assign a sentiment value to each piece of news using a combination of various Natural Language techniques. The sentiment for each sector is obtained by a ‘bottom-up’ aggregation of the sentiment of stocks belonging to each sector.
  • The index has outperformed the S&P by 5 percentage points a year over the past 5 years using this sector selection process. Since the S&P500’s peak in Dec ’21, the S&P 500 has dropped by 12% on a total return basis, in contrast Brain’s Sector Rotation index has risen by 6%!

Brain AI – Europe Top 10, 30 and 50 Stock Indices

  • Brain AI Europe indices leverage Brain’s proprietary Artificial Intelligence model to select stocks according to their predicted future returns on a monthly time horizon, from a universe made of the largest 600 European stocks.
  • The model implements a voting scheme of machine-learning classifiers that non-linearly combine a variety of inputs using a series of techniques aimed at mitigating the well-known overfitting problem for financial data with a low signal to noise ratio.
  • Some examples of inputs are: price and volume related metrics; market regime features such as volatility and market dispersion; the stock sector; and calendar features representing possible anomalies.
  • The AI Europe Top 10 has performed almost twice as well as the Eurostoxx 50 over a five year period. 
Back to all press releases →