Market indices are a way of measuring the performance of a group of assets or securities, in other words, a benchmark. Some famous examples are the Hang Seng Index, the FTSE 100 Index, and the S&P 500.

Index providers adjust their indices’ constituents periodically to make sure their indices reflect the actual performance of the market, which is known as Index Rebalancing. As many investors duplicate indices or buy related index funds to hedge market risk or apply a passive investment strategy, the changes in a market index can lead to abnormal stock price shifts or liquidity changes, which lead to investment opportunities.

However, it’s not that easy to predict the rebalancing:

1. Some indices have a committee to decide the rebalancing, which lacks transparency.
2. Although some indices have fixed rules (e.g., Pick the top 100 stocks with the highest market capital), the market itself is constantly changing.
3. Limited high-quality data is available.

This research project is centered on creating a predictive model for index rebalancing events utilizing machine learning techniques. The project will feed the historical data of index constituents into a machine-learning model to make predictions of the index’s future adjustments. Furthermore, a trading strategy will be developed based on the findings of the predicted results.