Machine learning for stock selection
Machine learning algorithms are increasingly used to solve problems in the financial markets. However, applications of the models for investment management and stock selection are still subject to debate with many challenges to overcome, such as model overfitting.
The Financial Analysts Journal explores this controversial topic further. Using the IHS Markit Research Signals factor library as a clean and robust feature set, the piece investigates the use of various machine learning models to forecast stock returns and build systematic portfolios. Options to leverage these tools are considered, while methods to overcome challenges such as overfitting, are also suggested. Will the results show that using a library of proven factors based on economic rationale as inputs to machine learning models produces strategies that generate statistically significant alpha and enhance stock selection?
To find out more read full article here
IHS Markit provides industry-leading data, software and technology platforms and managed services to tackle some of the most difficult challenges in financial markets. We help our customers better understand complicated markets, reduce risk, operate more efficiently and comply with financial regulation.
Follow Financial Services
- UK Flash PMI signals economic rebound as uncertainty lifts
- IHS Markit European GDP Nowcasts: Eurozone growth remains muted despite return to growth of Germany in first quarter
- US dividends to hit record high in 2020
- iBoxx Liquid Leveraged Loan TRS Reaches Record Volume
- LNG JKM Option Pricing - IHS Markit's Totem - Another First
- Seven key dividend forecasts for 2020
- Is BB the place to be?
- IHS Markit Fair Value Case Study
Our Global Carbon Index shows that the global weighted average price of carbon credits is $23.15. Since 23 January… https://t.co/JJVOXvP4xt