AI BriefingLG AI Research 384
Key point
LG AI Research's demand forecasting squad achieved award-winning results in 3 categories at the M6 forecasting competition held in New York.
Details
LG AI Research's Data Intelligence (DI) Lab demand forecasting squad participated in The Future of Forecasting and the M6 Competition Conference, held in New York in November 2023. The conference brought together experts in the field of Forecasting to share the latest insights from academia and industry, and LG AI Research was invited as an award winner in 3 categories at the M6 Competition.
The M6 Competition is a 48-week financial forecasting competition covering 100 listed assets. Participants compete in two tracks: the Forecasting challenge, which predicts the probability of which rank each asset will belong to based on asset returns, and Investment weight decisions, which determines portfolio investment weights.
The key evaluation metrics are as follows:
- RPS (Ranked Probability Score): Measures the accuracy of predicted rankings (closer to 0 is more accurate).
- IR (Information Ratio): Measures the stability and profitability of investment weight decisions (higher is better).
- OR (Overall): A comprehensive metric calculated as the arithmetic mean of RPS and IR.
Through this competition, LG AI Research shared various approaches ranging from AI-based demand forecasting to the financial sector, analyzed the gap between existing theory and actual markets, and explored future directions for forecasting technology.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Google Releases TimesFM-3 for Multivariate Time Series Forecasting
Demand Patterns Choose Demand Models: Agentic Demand Forecasting with Amazon Connect Decisions