Numerai: Portfolio Optimizer

What is Numerai?

  • Numerai is a quantitative hedge fund run by data scientists around the world. It holds a weekly (now daily) tournament to crowdsource machine learning models that predict the stock market.
  • Data scientists can download the data, build models, upload predictions and stake on model performance. The models are evaluated on the live market data and the best models are rewarded. See Numerai website for more details

What is this project about? Why does it matter?

  • Numerai participants can stake on their models / purchase models from community marketplace to earn rewards. A portfolio is a combination of models (in investment world -> assets) that you can stake on, each model with a different weights.

  • Balancing the risk and return trade-off in a portfolio is a challenging task, especially when there are a lot of highly correlated models.

  • This project aims to provide a simple, visualized tool to help Numerai participants to find the best portfolio, based on historical return and volatility.

What is a Portfolio Optimizer?

Portfolio Optimization originates from Harry Markowitz’s Model Portfolio Theory (1952). The core idea is to find a balance between the risk-return tradeoff, through constructing an efficient set of portfolios - Efficient Frontier. The concept is widely used in the financial industry, and it is also applicable to Numerai model staking. It takes into account the correlation between models, and historical return, to find the optimal weights of the models to build your portfolio.

Open In Colab

Link to repo: https://github.com/eses-wk/numerai-portfolio-opt

How does it work?

  • Step 1: Extract model historical performance result (Numerai Correlation / True Contribution) using Numerai API
  • Step 2: Use the optimizer to find the optimal weights of the models
  • Step 3: Review and select your optimal portfolio mix for model staking