用于金融深度强化学习的梯度降低卷积神经网络策略

Sina Montazeri, Haseebullah Jumakhan, Sonia Abrasiabian, Amir Mirzaeinia
{"title":"用于金融深度强化学习的梯度降低卷积神经网络策略","authors":"Sina Montazeri, Haseebullah Jumakhan, Sonia Abrasiabian, Amir Mirzaeinia","doi":"arxiv-2408.11859","DOIUrl":null,"url":null,"abstract":"Building on our prior explorations of convolutional neural networks (CNNs)\nfor financial data processing, this paper introduces two significant\nenhancements to refine our CNN model's predictive performance and robustness\nfor financial tabular data. Firstly, we integrate a normalization layer at the\ninput stage to ensure consistent feature scaling, addressing the issue of\ndisparate feature magnitudes that can skew the learning process. This\nmodification is hypothesized to aid in stabilizing the training dynamics and\nimproving the model's generalization across diverse financial datasets.\nSecondly, we employ a Gradient Reduction Architecture, where earlier layers are\nwider and subsequent layers are progressively narrower. This enhancement is\ndesigned to enable the model to capture more complex and subtle patterns within\nthe data, a crucial factor in accurately predicting financial outcomes. These\nadvancements directly respond to the limitations identified in previous\nstudies, where simpler models struggled with the complexity and variability\ninherent in financial applications. Initial tests confirm that these changes\nimprove accuracy and model stability, suggesting that deeper and more nuanced\nnetwork architectures can significantly benefit financial predictive tasks.\nThis paper details the implementation of these enhancements and evaluates their\nimpact on the model's performance in a controlled experimental setting.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"6 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Gradient Reduction Convolutional Neural Network Policy for Financial Deep Reinforcement Learning\",\"authors\":\"Sina Montazeri, Haseebullah Jumakhan, Sonia Abrasiabian, Amir Mirzaeinia\",\"doi\":\"arxiv-2408.11859\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Building on our prior explorations of convolutional neural networks (CNNs)\\nfor financial data processing, this paper introduces two significant\\nenhancements to refine our CNN model's predictive performance and robustness\\nfor financial tabular data. Firstly, we integrate a normalization layer at the\\ninput stage to ensure consistent feature scaling, addressing the issue of\\ndisparate feature magnitudes that can skew the learning process. This\\nmodification is hypothesized to aid in stabilizing the training dynamics and\\nimproving the model's generalization across diverse financial datasets.\\nSecondly, we employ a Gradient Reduction Architecture, where earlier layers are\\nwider and subsequent layers are progressively narrower. This enhancement is\\ndesigned to enable the model to capture more complex and subtle patterns within\\nthe data, a crucial factor in accurately predicting financial outcomes. These\\nadvancements directly respond to the limitations identified in previous\\nstudies, where simpler models struggled with the complexity and variability\\ninherent in financial applications. Initial tests confirm that these changes\\nimprove accuracy and model stability, suggesting that deeper and more nuanced\\nnetwork architectures can significantly benefit financial predictive tasks.\\nThis paper details the implementation of these enhancements and evaluates their\\nimpact on the model's performance in a controlled experimental setting.\",\"PeriodicalId\":501294,\"journal\":{\"name\":\"arXiv - QuantFin - Computational Finance\",\"volume\":\"6 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - QuantFin - Computational Finance\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.11859\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - QuantFin - Computational Finance","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.11859","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

基于我们之前对用于金融数据处理的卷积神经网络(CNN)的探索,本文引入了两项重大改进,以完善 CNN 模型的预测性能和对金融表格数据的鲁棒性。首先,我们在输入阶段集成了一个归一化层,以确保一致的特征缩放,从而解决可能会影响学习过程的不同特征量级的问题。这一修改被认为有助于稳定训练动态,提高模型在不同金融数据集上的泛化能力。其次,我们采用了梯度缩减架构,即前面的层更宽,后面的层逐渐变窄。这一改进旨在使模型能够捕捉数据中更复杂、更微妙的模式,这是准确预测金融结果的关键因素。在以往的研究中,较简单的模型难以应对金融应用中固有的复杂性和可变性,而这些改进直接应对了这些局限性。初步测试证实,这些变化提高了准确性和模型的稳定性,表明更深入、更细致的网络架构能显著提高金融预测任务的效率。本文详细介绍了这些增强功能的实现过程,并在受控实验环境中评估了它们对模型性能的影响。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Gradient Reduction Convolutional Neural Network Policy for Financial Deep Reinforcement Learning
Building on our prior explorations of convolutional neural networks (CNNs) for financial data processing, this paper introduces two significant enhancements to refine our CNN model's predictive performance and robustness for financial tabular data. Firstly, we integrate a normalization layer at the input stage to ensure consistent feature scaling, addressing the issue of disparate feature magnitudes that can skew the learning process. This modification is hypothesized to aid in stabilizing the training dynamics and improving the model's generalization across diverse financial datasets. Secondly, we employ a Gradient Reduction Architecture, where earlier layers are wider and subsequent layers are progressively narrower. This enhancement is designed to enable the model to capture more complex and subtle patterns within the data, a crucial factor in accurately predicting financial outcomes. These advancements directly respond to the limitations identified in previous studies, where simpler models struggled with the complexity and variability inherent in financial applications. Initial tests confirm that these changes improve accuracy and model stability, suggesting that deeper and more nuanced network architectures can significantly benefit financial predictive tasks. This paper details the implementation of these enhancements and evaluates their impact on the model's performance in a controlled experimental setting.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信