arXiv - QuantFin - Portfolio Management最新文献_第2页

Investment strategies based on forecasts are (almost) useless 基于预测的投资战略（几乎）毫无用处

arXiv - QuantFin - Portfolio Management Pub Date : 2024-08-03 DOI: arxiv-2408.01772

Michael Weba

引用次数: 0

Risk management in multi-objective portfolio optimization under uncertainty 不确定情况下多目标组合优化的风险管理

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-29 DOI: arxiv-2407.19936

Yannick Becker, Pascal Halffmann, Anita Schöbel

引用次数: 0

PO-QA: A Framework for Portfolio Optimization using Quantum Algorithms PO-QA：利用量子算法优化投资组合的框架

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-29 DOI: arxiv-2407.19857

Kamila Zaman, Alberto Marchisio, Muhammad Kashif, Muhammad Shafique

{"title":"PO-QA: A Framework for Portfolio Optimization using Quantum Algorithms","authors":"Kamila Zaman, Alberto Marchisio, Muhammad Kashif, Muhammad Shafique","doi":"arxiv-2407.19857","DOIUrl":"https://doi.org/arxiv-2407.19857","url":null,"abstract":"Portfolio Optimization (PO) is a financial problem aiming to maximize the net\u0000gains while minimizing the risks in a given investment portfolio. The novelty\u0000of Quantum algorithms lies in their acclaimed potential and capability to solve\u0000complex problems given the underlying Quantum Computing (QC) infrastructure.\u0000Utilizing QC's applicable strengths to the finance industry's problems, such as\u0000PO, allows us to solve these problems using quantum-based algorithms such as\u0000Variational Quantum Eigensolver (VQE) and Quantum Approximate Optimization\u0000Algorithm (QAOA). While the Quantum potential for finance is highly impactful,\u0000the architecture and composition of the quantum circuits have not yet been\u0000properly defined as robust financial frameworks/algorithms as state of the art\u0000in present literature for research and design development purposes. In this\u0000work, we propose a novel scalable framework, denoted PO-QA, to systematically\u0000investigate the variation of quantum parameters (such as rotation blocks,\u0000repetitions, and entanglement types) to observe their subtle effect on the\u0000overall performance. In our paper, the performance is measured and dictated by\u0000convergence to similar ground-state energy values for resultant optimal\u0000solutions by each algorithm variation set for QAOA and VQE to the exact\u0000eigensolver (classical solution). Our results provide effective insights into\u0000comprehending PO from the lens of Quantum Machine Learning in terms of\u0000convergence to the classical solution, which is used as a benchmark. This study\u0000paves the way for identifying efficient configurations of quantum circuits for\u0000solving PO and unveiling their inherent inter-relationships.","PeriodicalId":501045,"journal":{"name":"arXiv - QuantFin - Portfolio Management","volume":"89 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141872589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimal retirement in presence of stochastic labor income: a free boundary approach in presence of an incomplete market 随机劳动收入条件下的最优退休：不完全市场条件下的自由边界法

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-27 DOI: arxiv-2407.19190

Daniele Marazzina

引用次数: 0

Fine-Tuning Large Language Models for Stock Return Prediction Using Newsflow 利用新闻流微调用于股票回报预测的大型语言模型

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-25 DOI: arxiv-2407.18103

Tian Guo, Emmanuel Hauptmann

{"title":"Fine-Tuning Large Language Models for Stock Return Prediction Using Newsflow","authors":"Tian Guo, Emmanuel Hauptmann","doi":"arxiv-2407.18103","DOIUrl":"https://doi.org/arxiv-2407.18103","url":null,"abstract":"Large language models (LLMs) and their fine-tuning techniques have\u0000demonstrated superior performance in various language understanding and\u0000generation tasks. This paper explores fine-tuning LLMs for stock return\u0000forecasting with financial newsflow. In quantitative investing, return\u0000forecasting is fundamental for subsequent tasks like stock picking, portfolio\u0000optimization, etc. We formulate the model to include text representation and\u0000forecasting modules. We propose to compare the encoder-only and decoder-only\u0000LLMs, considering they generate text representations in distinct ways. The\u0000impact of these different representations on forecasting performance remains an\u0000open question. Meanwhile, we compare two simple methods of integrating LLMs'\u0000token-level representations into the forecasting module. The experiments on\u0000real news and investment universes reveal that: (1) aggregated representations\u0000from LLMs' token-level embeddings generally produce return predictions that\u0000enhance the performance of long-only and long-short portfolios; (2) in the\u0000relatively large investment universe, the decoder LLMs-based prediction model\u0000leads to stronger portfolios, whereas in the small universes, there are no\u0000consistent winners. Among the three LLMs studied (DeBERTa, Mistral, Llama),\u0000Mistral performs more robustly across different universes; (3) return\u0000predictions derived from LLMs' text representations are a strong signal for\u0000portfolio construction, outperforming conventional sentiment scores.","PeriodicalId":501045,"journal":{"name":"arXiv - QuantFin - Portfolio Management","volume":"61 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Hopfield Networks for Asset Allocation 用于资产配置的 Hopfield 网络

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-24 DOI: arxiv-2407.17645

Carlo Nicolini, Monisha Gopalan, Jacopo Staiano, Bruno Lepri

引用次数: 0

Multi-Industry Simplex 2.0 : Temporally-Evolving Probabilistic Industry Classification 多产业 Simplex 2.0：时变概率产业分类

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-23 DOI: arxiv-2407.16437

Maksim Papenkov

{"title":"Multi-Industry Simplex 2.0 : Temporally-Evolving Probabilistic Industry Classification","authors":"Maksim Papenkov","doi":"arxiv-2407.16437","DOIUrl":"https://doi.org/arxiv-2407.16437","url":null,"abstract":"Accurate industry classification is critical for many areas of portfolio\u0000management, yet the traditional single-industry framework of the Global\u0000Industry Classification Standard (GICS) struggles to comprehensively represent\u0000risk for highly diversified multi-sector conglomerates like Amazon. Previously,\u0000we introduced the Multi-Industry Simplex (MIS), a probabilistic extension of\u0000GICS that utilizes topic modeling, a natural language processing approach.\u0000Although our initial version, MIS-1, was able to improve upon GICS by providing\u0000multi-industry representations, it relied on an overly simple architecture that\u0000required prior knowledge about the number of industries and relied on the\u0000unrealistic assumption that industries are uncorrelated and independent over\u0000time. We improve upon this model with MIS-2, which addresses three key\u0000limitations of MIS-1 : we utilize Bayesian Non-Parametrics to automatically\u0000infer the number of industries from data, we employ Markov Updating to account\u0000for industries that change over time, and we adjust for correlated and\u0000hierarchical industries allowing for both broad and niche industries (similar\u0000to GICS). Further, we provide an out-of-sample test directly comparing MIS-2\u0000and GICS on the basis of future correlation prediction, where we find evidence\u0000that MIS-2 provides a measurable improvement over GICS. MIS-2 provides\u0000portfolio managers with a more robust tool for industry classification,\u0000empowering them to more effectively identify and manage risk, particularly\u0000around multi-sector conglomerates in a rapidly evolving market in which new\u0000industries periodically emerge.","PeriodicalId":501045,"journal":{"name":"arXiv - QuantFin - Portfolio Management","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771836","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Explainable Post hoc Portfolio Management Financial Policy of a Deep Reinforcement Learning agent 深度强化学习代理的可解释临时投资组合管理财务政策

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-19 DOI: arxiv-2407.14486

Alejandra de la Rica Escudero, Eduardo C. Garrido-Merchan, Maria Coronado-Vaca

{"title":"Explainable Post hoc Portfolio Management Financial Policy of a Deep Reinforcement Learning agent","authors":"Alejandra de la Rica Escudero, Eduardo C. Garrido-Merchan, Maria Coronado-Vaca","doi":"arxiv-2407.14486","DOIUrl":"https://doi.org/arxiv-2407.14486","url":null,"abstract":"Financial portfolio management investment policies computed quantitatively by\u0000modern portfolio theory techniques like the Markowitz model rely on a set on\u0000assumptions that are not supported by data in high volatility markets. Hence,\u0000quantitative researchers are looking for alternative models to tackle this\u0000problem. Concretely, portfolio management is a problem that has been\u0000successfully addressed recently by Deep Reinforcement Learning (DRL)\u0000approaches. In particular, DRL algorithms train an agent by estimating the\u0000distribution of the expected reward of every action performed by an agent given\u0000any financial state in a simulator. However, these methods rely on Deep Neural\u0000Networks model to represent such a distribution, that although they are\u0000universal approximator models, they cannot explain its behaviour, given by a\u0000set of parameters that are not interpretable. Critically, financial investors\u0000policies require predictions to be interpretable, so DRL agents are not suited\u0000to follow a particular policy or explain their actions. In this work, we\u0000developed a novel Explainable Deep Reinforcement Learning (XDRL) approach for\u0000portfolio management, integrating the Proximal Policy Optimization (PPO) with\u0000the model agnostic explainable techniques of feature importance, SHAP and LIME\u0000to enhance transparency in prediction time. By executing our methodology, we\u0000can interpret in prediction time the actions of the agent to assess whether\u0000they follow the requisites of an investment policy or to assess the risk of\u0000following the agent suggestions. To the best of our knowledge, our proposed\u0000approach is the first explainable post hoc portfolio management financial\u0000policy of a DRL agent. We empirically illustrate our methodology by\u0000successfully identifying key features influencing investment decisions, which\u0000demonstrate the ability to explain the agent actions in prediction time.","PeriodicalId":501045,"journal":{"name":"arXiv - QuantFin - Portfolio Management","volume":"77 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141745023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Longitudinal market structure detection using a dynamic modularity-spectral algorithm 使用动态模块化光谱算法进行纵向市场结构检测

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-05 DOI: arxiv-2407.04500

Philipp Wirth, Francesca Medda, Thomas Schröder

{"title":"Longitudinal market structure detection using a dynamic modularity-spectral algorithm","authors":"Philipp Wirth, Francesca Medda, Thomas Schröder","doi":"arxiv-2407.04500","DOIUrl":"https://doi.org/arxiv-2407.04500","url":null,"abstract":"In this paper, we introduce the Dynamic Modularity-Spectral Algorithm\u0000(DynMSA), a novel approach to identify clusters of stocks with high\u0000intra-cluster correlations and low inter-cluster correlations by combining\u0000Random Matrix Theory with modularity optimisation and spectral clustering. The\u0000primary objective is to uncover hidden market structures and find diversifiers\u0000based on return correlations, thereby achieving a more effective risk-reducing\u0000portfolio allocation. We applied DynMSA to constituents of the S&P 500 and\u0000compared the results to sector- and market-based benchmarks. Besides the\u0000conception of this algorithm, our contributions further include implementing a\u0000sector-based calibration for modularity optimisation and a correlation-based\u0000distance function for spectral clustering. Testing revealed that DynMSA\u0000outperforms baseline models in intra- and inter-cluster correlation\u0000differences, particularly over medium-term correlation look-backs. It also\u0000identifies stable clusters and detects regime changes due to exogenous shocks,\u0000such as the COVID-19 pandemic. Portfolios constructed using our clusters showed\u0000higher Sortino and Sharpe ratios, lower downside volatility, reduced maximum\u0000drawdown and higher annualised returns compared to an equally weighted market\u0000benchmark.","PeriodicalId":501045,"journal":{"name":"arXiv - QuantFin - Portfolio Management","volume":"57 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141576657","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Robust optimal investment and consumption strategies with portfolio constraints and stochastic environment 具有投资组合约束和随机环境的稳健优化投资和消费策略

arXiv - QuantFin - Portfolio Management Pub Date : 2024-07-03 DOI: arxiv-2407.02831

Len Patrick Dominic M. Garces, Yang Shen

{"title":"Robust optimal investment and consumption strategies with portfolio constraints and stochastic environment","authors":"Len Patrick Dominic M. Garces, Yang Shen","doi":"arxiv-2407.02831","DOIUrl":"https://doi.org/arxiv-2407.02831","url":null,"abstract":"We investigate a continuous-time investment-consumption problem with model\u0000uncertainty in a general diffusion-based market with random model coefficients.\u0000We assume that a power utility investor is ambiguity-averse, with the\u0000preference to robustness captured by the homothetic multiplier robust\u0000specification, and the investor's investment and consumption strategies are\u0000constrained to closed convex sets. To solve this constrained robust control\u0000problem, we employ the stochastic Hamilton-Jacobi-Bellman-Isaacs equations,\u0000backward stochastic differential equations, and bounded mean oscillation\u0000martingale theory. Furthermore, we show the investor incurs (non-negative)\u0000utility loss, i.e. the loss in welfare, if model uncertainty is ignored. When\u0000the model coefficients are deterministic, we establish formally the\u0000relationship between the investor's robustness preference and the robust\u0000optimal investment-consumption strategy and the value function, and the impact\u0000of investment and consumption constraints on the investor's robust optimal\u0000investment-consumption strategy and value function. Extensive numerical\u0000experiments highlight the significant impact of ambiguity aversion, consumption\u0000and investment constraints, on the investor's robust optimal\u0000investment-consumption strategy, utility loss, and value function. Key findings\u0000include: 1) short-selling restriction always reduces the investor's utility\u0000loss when model uncertainty is ignored; 2) the effect of consumption\u0000constraints on utility loss is more delicate and relies on the investor's risk\u0000aversion level.","PeriodicalId":501045,"journal":{"name":"arXiv - QuantFin - Portfolio Management","volume":"12 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141548543","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0