DSPO: An End-to-End Framework for Direct Sorted Portfolio Construction
DSPO is an end-to-end framework for constructing sorted portfolios using deep learning on multi-frequency stock data.
What it examines
The paper introduces Direct Sorted Portfolio Optimization (DSPO), an end-to-end framework for constructing characteristic-sorted portfolios directly from raw stock data. It addresses limitations in traditional methods by integrating multi-frequency data and optimizing directly for portfolio sorting using a neural network architecture.
What it concludes
The study concludes that DSPO is a robust and efficient framework for portfolio construction, leveraging diverse market data. Potential applications include quantitative investment strategies and financial market analysis. Future research should focus on integrating risk management strategies to enhance DSPO's adaptability to varying market conditions.
Evidence objects
The study concludes that DSPO is a robust and efficient framework for portfolio construction, leveraging diverse market data. Potential applications include quantitative investment strategies and financial market analysis. Future research should focus on integrating risk management strategies to enhance DSPO's adaptability to varying market conditions.
key_findings bullet 1 · key_findings · validation V0
Raw abstract and provenance
Abstract: In quantitative investment, constructing characteristic-sorted portfolios is a crucial strategy for asset allocation. Traditional methods transform raw stock data of varying frequencies into predictive characteristic factors for asset sorting, often requiring extensive manual design and misalignment between prediction and optimization goals. To address these challenges, we introduce Direct Sorted… ▽ More In quantitative investment, constructing characteristic-sorted portfolios is a crucial strategy for asset allocation. Traditional methods transform raw stock data of varying frequencies into predictive characteristic factors for asset sorting, often requiring extensive manual design and misalignment between prediction and optimization goals. To address these challenges, we introduce Direct Sorted Portfolio Optimization (DSPO), an innovative end-to-end framework that efficiently processes raw stock data to construct sorted portfolios directly. DSPO's neural network architecture seamlessly transitions stock data from input to output while effectively modeling the intra-dependency of time-steps and inter-dependency among all tradable stocks. Additionally, we incorporate a novel Monotonical Logistic Regression loss, which directly maximizes the likelihood of constructing optimal sorted portfolios. To the best of our knowledge, DSPO is the first method capable of handling market cross-sections with thousands of tradable stocks fully end-to-end from raw multi-frequency data. Empirical results demonstrate DSPO's effectiveness, yielding a RankIC of 10.12\% and an accumulated return of 121.94\% on the New York Stock Exchange in 2023-2024, and a RankIC of 9.11\% with a return of 108.74\% in other markets during 2021-2022. △ Less
Source row: 642 · abstract type: unknown