---
title: Machine Learning for Portfolio Selection
url: https://www.ml-quant.com/papers/ssrn/4669599/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: SSRN 4669599
source_url: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4669599
featured: 2023-12-20
citations: unknown
topic: Portfolio & Allocation
---


# Machine Learning for Portfolio Selection

A new performance ratio is created to address the limitations of the Sharpe ratio under non-Gaussian returns and systemic risk, showing improved portfolio selection performance in terms of profitability and risk reduction.

- Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4669599
- Identifier: SSRN 4669599
- Released: 2023-08-11
- First featured: Quant Letter No. 30 (2023-12-20): https://www.ml-quant.com/issues/2023-12-20/
- Citations (Semantic Scholar): not tracked
- Published in: not yet
- Topic: Portfolio & Allocation

## Related

- [Partial Information in a Mean‐Variance Portfolio Selection Game](https://www.ml-quant.com/papers/arxiv/2312.04045/): The article investigates how limited information affects investors' wealth and systemic risk, using a model where investors adjust their strategies based on their wealth compared to others.
- [Set risk measures](https://www.ml-quant.com/papers/arxiv/2407.18687/): The study presents set risk measures, which extend traditional risk measures to sets of random variables, and establishes an axiom scheme for them, demonstrating their use in systemic risk, portfolio optimization, and decision-making under uncertainty.
- [Systemic Risk from Overlapping Portfolios](https://www.ml-quant.com/papers/repec/eee-finana-v-97-y-2025-i-c-s1057521924007269/): A portfolio optimization framework accounting for systemic and individual risk reveals potential inefficiencies in portfolio structures, indicating a risk trade-off.
- [Sparse Portfolio Optimization via a Novel Fractional Regularization](https://www.ml-quant.com/papers/ssrn/4666990/): The paper introduces a new L1L2 regularized sparse portfolio optimization model using the ADMM method, and discusses an extension of the model to include a more general L1Lq regularization.
- [Outperforming Equal Weighting](https://www.ml-quant.com/papers/ssrn/4669267/): The article suggests that an equally-weighted stock portfolio can be improved by avoiding negative exposure to certain factor anomalies, while keeping the portfolio construction process simple.
- [A return-diversification approach to portfolio selection](https://www.ml-quant.com/papers/arxiv/2312.09707/): The article suggests a dual-objective model for portfolio selection that optimizes both diversification and expected return, outperforming strategies focused only on diversification or risk-return.