We present a new deep primal-dual backward stochastic differential equation framework based on stopping time iteration to solve optimal stopping problems. A novel loss function is proposed to learn the conditional expectation, which consists of subnetwork parameterization of a continuation value and spatial gradients from present up to the stopping time. Notable features of the method include: (i) The martingale part in the loss function reduces the variance of stochastic gradients, which facilitates the training of the neural networks as well as alleviates the error propagation of value function approximation; (ii) this martingale approximates the martingale in the Doob-Meyer decomposition, and thus leads to a true upper bound for the optimal value in a non-nested Monte Carlo way. We test the proposed method in American option pricing problems, where the spatial gradient network yields the hedging ratio directly.
{"title":"A deep primal-dual BSDE method for optimal stopping problems","authors":"Jiefei Yang, Guanglian Li","doi":"arxiv-2409.06937","DOIUrl":"https://doi.org/arxiv-2409.06937","url":null,"abstract":"We present a new deep primal-dual backward stochastic differential equation\u0000framework based on stopping time iteration to solve optimal stopping problems.\u0000A novel loss function is proposed to learn the conditional expectation, which\u0000consists of subnetwork parameterization of a continuation value and spatial\u0000gradients from present up to the stopping time. Notable features of the method\u0000include: (i) The martingale part in the loss function reduces the variance of\u0000stochastic gradients, which facilitates the training of the neural networks as\u0000well as alleviates the error propagation of value function approximation; (ii)\u0000this martingale approximates the martingale in the Doob-Meyer decomposition,\u0000and thus leads to a true upper bound for the optimal value in a non-nested\u0000Monte Carlo way. We test the proposed method in American option pricing\u0000problems, where the spatial gradient network yields the hedging ratio directly.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"12 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208118","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}
The paper presents a Bayesian framework for the calibration of financial models using neural stochastic differential equations (neural SDEs). The method is based on the specification of a prior distribution on the neural network weights and an adequately chosen likelihood function. The resulting posterior distribution can be seen as a mixture of different classical neural SDE models yielding robust bounds on the implied volatility surface. Both, historical financial time series data and option price data are taken into consideration, which necessitates a methodology to learn the change of measure between the risk-neutral and the historical measure. The key ingredient for a robust numerical optimization of the neural networks is to apply a Langevin-type algorithm, commonly used in the Bayesian approaches to draw posterior samples.
{"title":"Robust financial calibration: a Bayesian approach for neural SDEs","authors":"Christa Cuchiero, Eva Flonner, Kevin Kurt","doi":"arxiv-2409.06551","DOIUrl":"https://doi.org/arxiv-2409.06551","url":null,"abstract":"The paper presents a Bayesian framework for the calibration of financial\u0000models using neural stochastic differential equations (neural SDEs). The method\u0000is based on the specification of a prior distribution on the neural network\u0000weights and an adequately chosen likelihood function. The resulting posterior\u0000distribution can be seen as a mixture of different classical neural SDE models\u0000yielding robust bounds on the implied volatility surface. Both, historical\u0000financial time series data and option price data are taken into consideration,\u0000which necessitates a methodology to learn the change of measure between the\u0000risk-neutral and the historical measure. The key ingredient for a robust\u0000numerical optimization of the neural networks is to apply a Langevin-type\u0000algorithm, commonly used in the Bayesian approaches to draw posterior samples.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"252 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208119","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}
It is widely acknowledged that extracting market sentiments from news data benefits market predictions. However, existing methods of using financial sentiments remain simplistic, relying on equal-weight and static aggregation to manage sentiments from multiple news items. This leads to a critical issue termed ``Aggregated Sentiment Homogenization'', which has been explored through our analysis of a large financial news dataset from industry practice. This phenomenon occurs when aggregating numerous sentiments, causing representations to converge towards the mean values of sentiment distributions and thereby smoothing out unique and important information. Consequently, the aggregated sentiment representations lose much predictive value of news data. To address this problem, we introduce the Market Attention-weighted News Aggregation Network (MANA-Net), a novel method that leverages a dynamic market-news attention mechanism to aggregate news sentiments for market prediction. MANA-Net learns the relevance of news sentiments to price changes and assigns varying weights to individual news items. By integrating the news aggregation step into the networks for market prediction, MANA-Net allows for trainable sentiment representations that are optimized directly for prediction. We evaluate MANA-Net using the S&P 500 and NASDAQ 100 indices, along with financial news spanning from 2003 to 2018. Experimental results demonstrate that MANA-Net outperforms various recent market prediction methods, enhancing Profit & Loss by 1.1% and the daily Sharpe ratio by 0.252.
{"title":"MANA-Net: Mitigating Aggregated Sentiment Homogenization with News Weighting for Enhanced Market Prediction","authors":"Mengyu Wang, Tiejun Ma","doi":"arxiv-2409.05698","DOIUrl":"https://doi.org/arxiv-2409.05698","url":null,"abstract":"It is widely acknowledged that extracting market sentiments from news data\u0000benefits market predictions. However, existing methods of using financial\u0000sentiments remain simplistic, relying on equal-weight and static aggregation to\u0000manage sentiments from multiple news items. This leads to a critical issue\u0000termed ``Aggregated Sentiment Homogenization'', which has been explored through\u0000our analysis of a large financial news dataset from industry practice. This\u0000phenomenon occurs when aggregating numerous sentiments, causing representations\u0000to converge towards the mean values of sentiment distributions and thereby\u0000smoothing out unique and important information. Consequently, the aggregated\u0000sentiment representations lose much predictive value of news data. To address\u0000this problem, we introduce the Market Attention-weighted News Aggregation\u0000Network (MANA-Net), a novel method that leverages a dynamic market-news\u0000attention mechanism to aggregate news sentiments for market prediction.\u0000MANA-Net learns the relevance of news sentiments to price changes and assigns\u0000varying weights to individual news items. By integrating the news aggregation\u0000step into the networks for market prediction, MANA-Net allows for trainable\u0000sentiment representations that are optimized directly for prediction. We\u0000evaluate MANA-Net using the S&P 500 and NASDAQ 100 indices, along with\u0000financial news spanning from 2003 to 2018. Experimental results demonstrate\u0000that MANA-Net outperforms various recent market prediction methods, enhancing\u0000Profit & Loss by 1.1% and the daily Sharpe ratio by 0.252.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142226610","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}
The goal of alpha factor mining is to discover indicative signals of investment opportunities from the historical financial market data of assets. Deep learning based alpha factor mining methods have shown to be powerful, which, however, lack of the interpretability, making them unacceptable in the risk-sensitive real markets. Alpha factors in formulaic forms are more interpretable and therefore favored by market participants, while the search space is complex and powerful explorative methods are urged. Recently, a promising framework is proposed for generating formulaic alpha factors using deep reinforcement learning, and quickly gained research focuses from both academia and industries. This paper first argues that the originally employed policy training method, i.e., Proximal Policy Optimization (PPO), faces several important issues in the context of alpha factors mining, making it ineffective to explore the search space of the formula. Herein, a novel reinforcement learning based on the well-known REINFORCE algorithm is proposed. Given that the underlying state transition function adheres to the Dirac distribution, the Markov Decision Process within this framework exhibit minimal environmental variability, making REINFORCE algorithm more appropriate than PPO. A new dedicated baseline is designed to theoretically reduce the commonly suffered high variance of REINFORCE. Moreover, the information ratio is introduced as a reward shaping mechanism to encourage the generation of steady alpha factors that can better adapt to changes in market volatility. Experimental evaluations on various real assets data show that the proposed algorithm can increase the correlation with asset returns by 3.83%, and a stronger ability to obtain excess returns compared to the latest alpha factors mining methods, which meets the theoretical results well.
{"title":"QuantFactor REINFORCE: Mining Steady Formulaic Alpha Factors with Variance-bounded REINFORCE","authors":"Junjie Zhao, Chengxi Zhang, Min Qin, Peng Yang","doi":"arxiv-2409.05144","DOIUrl":"https://doi.org/arxiv-2409.05144","url":null,"abstract":"The goal of alpha factor mining is to discover indicative signals of\u0000investment opportunities from the historical financial market data of assets.\u0000Deep learning based alpha factor mining methods have shown to be powerful,\u0000which, however, lack of the interpretability, making them unacceptable in the\u0000risk-sensitive real markets. Alpha factors in formulaic forms are more\u0000interpretable and therefore favored by market participants, while the search\u0000space is complex and powerful explorative methods are urged. Recently, a\u0000promising framework is proposed for generating formulaic alpha factors using\u0000deep reinforcement learning, and quickly gained research focuses from both\u0000academia and industries. This paper first argues that the originally employed\u0000policy training method, i.e., Proximal Policy Optimization (PPO), faces several\u0000important issues in the context of alpha factors mining, making it ineffective\u0000to explore the search space of the formula. Herein, a novel reinforcement\u0000learning based on the well-known REINFORCE algorithm is proposed. Given that\u0000the underlying state transition function adheres to the Dirac distribution, the\u0000Markov Decision Process within this framework exhibit minimal environmental\u0000variability, making REINFORCE algorithm more appropriate than PPO. A new\u0000dedicated baseline is designed to theoretically reduce the commonly suffered\u0000high variance of REINFORCE. Moreover, the information ratio is introduced as a\u0000reward shaping mechanism to encourage the generation of steady alpha factors\u0000that can better adapt to changes in market volatility. Experimental evaluations\u0000on various real assets data show that the proposed algorithm can increase the\u0000correlation with asset returns by 3.83%, and a stronger ability to obtain\u0000excess returns compared to the latest alpha factors mining methods, which meets\u0000the theoretical results well.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208120","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}
Asim Ghosh, Soumyajyoti Biswas, Bikas K. Chakrabarti
We study the fluctuations, particularly the inequality of fluctuations, in cryptocurrency prices over the last ten years. We calculate the inequality in the price fluctuations through different measures, such as the Gini and Kolkata indices, and also the $Q$ factor (given by the ratio between the highest value and the average value) of these fluctuations. We compare the results with the equivalent quantities in some of the more prominent national currencies and see that while the fluctuations (or inequalities in such fluctuations) for cryptocurrencies were initially significantly higher than national currencies, over time the fluctuation levels of cryptocurrencies tend towards the levels characteristic of national currencies. We also compare similar quantities for a few prominent stock prices.
{"title":"Signature of maturity in cryptocurrency volatility","authors":"Asim Ghosh, Soumyajyoti Biswas, Bikas K. Chakrabarti","doi":"arxiv-2409.03676","DOIUrl":"https://doi.org/arxiv-2409.03676","url":null,"abstract":"We study the fluctuations, particularly the inequality of fluctuations, in\u0000cryptocurrency prices over the last ten years. We calculate the inequality in\u0000the price fluctuations through different measures, such as the Gini and Kolkata\u0000indices, and also the $Q$ factor (given by the ratio between the highest value\u0000and the average value) of these fluctuations. We compare the results with the\u0000equivalent quantities in some of the more prominent national currencies and see\u0000that while the fluctuations (or inequalities in such fluctuations) for\u0000cryptocurrencies were initially significantly higher than national currencies,\u0000over time the fluctuation levels of cryptocurrencies tend towards the levels\u0000characteristic of national currencies. We also compare similar quantities for a\u0000few prominent stock prices.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"6 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208121","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}
This study investigates the application of machine learning algorithms, particularly in the context of pricing American options using Monte Carlo simulations. Traditional models, such as the Black-Scholes-Merton framework, often fail to adequately address the complexities of American options, which include the ability for early exercise and non-linear payoff structures. By leveraging Monte Carlo methods in conjunction Least Square Method machine learning was used. This research aims to improve the accuracy and efficiency of option pricing. The study evaluates several machine learning models, including neural networks and decision trees, highlighting their potential to outperform traditional approaches. The results from applying machine learning algorithm in LSM indicate that integrating machine learning with Monte Carlo simulations can enhance pricing accuracy and provide more robust predictions, offering significant insights into quantitative finance by merging classical financial theories with modern computational techniques. The dataset was split into features and the target variable representing bid prices, with an 80-20 train-validation split. LSTM and GRU models were constructed using TensorFlow's Keras API, each with four hidden layers of 200 neurons and an output layer for bid price prediction, optimized with the Adam optimizer and MSE loss function. The GRU model outperformed the LSTM model across all evaluated metrics, demonstrating lower mean absolute error, mean squared error, and root mean squared error, along with greater stability and efficiency in training.
本研究探讨了机器学习算法的应用,尤其是在使用蒙特卡洛模拟法为美式期权定价时的应用。传统模型,如布莱克-斯科尔斯-默顿框架,往往无法充分解决美式期权的复杂性,其中包括提前行使能力和非线性报酬结构。本研究将蒙特卡罗方法与最小二乘法机器学习相结合。这项研究旨在提高期权定价的准确性和效率。研究评估了几种机器学习模型,包括神经网络和决策树,突出了它们优于传统方法的潜力。将机器学习算法应用于LSM 的结果表明,将机器学习与蒙特卡罗模拟相结合可以提高定价的准确性,并提供更稳健的预测,通过将经典金融理论与现代计算技术相结合,为定量金融学提供了重要见解。数据集被分为特征和代表投标价格的目标变量,训练-验证的比例为 80-20。使用 TensorFlow 的 Keras API 构建了 LSTM 和 GRU 模型,每个模型都有四个由 200 个神经元组成的隐藏层和一个禁止价格预测的输出层,并使用 Adam 优化器和 MSE 损失函数进行了优化。GRU 模型在所有评估指标上都优于 LSTM 模型,表现出更低的平均绝对误差、均方误差和均方根误差,以及更高的稳定性和训练效率。
{"title":"Pricing American Options using Machine Learning Algorithms","authors":"Prudence Djagba, Callixte Ndizihiwe","doi":"arxiv-2409.03204","DOIUrl":"https://doi.org/arxiv-2409.03204","url":null,"abstract":"This study investigates the application of machine learning algorithms,\u0000particularly in the context of pricing American options using Monte Carlo\u0000simulations. Traditional models, such as the Black-Scholes-Merton framework,\u0000often fail to adequately address the complexities of American options, which\u0000include the ability for early exercise and non-linear payoff structures. By\u0000leveraging Monte Carlo methods in conjunction Least Square Method machine\u0000learning was used. This research aims to improve the accuracy and efficiency of\u0000option pricing. The study evaluates several machine learning models, including\u0000neural networks and decision trees, highlighting their potential to outperform\u0000traditional approaches. The results from applying machine learning algorithm in\u0000LSM indicate that integrating machine learning with Monte Carlo simulations can\u0000enhance pricing accuracy and provide more robust predictions, offering\u0000significant insights into quantitative finance by merging classical financial\u0000theories with modern computational techniques. The dataset was split into\u0000features and the target variable representing bid prices, with an 80-20\u0000train-validation split. LSTM and GRU models were constructed using TensorFlow's\u0000Keras API, each with four hidden layers of 200 neurons and an output layer for\u0000bid price prediction, optimized with the Adam optimizer and MSE loss function.\u0000The GRU model outperformed the LSTM model across all evaluated metrics,\u0000demonstrating lower mean absolute error, mean squared error, and root mean\u0000squared error, along with greater stability and efficiency in training.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"22 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208122","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}
Sandy Chen, Leqi Zeng, Abhinav Raghunathan, Flora Huang, Terrence C. Kim
Large Language Models (LLMs) research in the financial domain is particularly complex due to the sheer number of approaches proposed in literature. Retrieval-Augmented Generation (RAG) has emerged as one of the leading methods in the sector due to its inherent groundedness and data source variability. In this work, we introduce a RAG framework called Mixture of Agents (MoA) and demonstrate its viability as a practical, customizable, and highly effective approach for scaling RAG applications. MoA is essentially a layered network of individually customized small language models (Hoffmann et al., 2022) collaborating to answer questions and extract information. While there are many theoretical propositions for such an architecture and even a few libraries for generally applying the structure in practice, there are limited documented studies evaluating the potential of this framework considering real business constraints such as cost and speed. We find that the MoA framework, consisting of small language models (Hoffmann et al., 2022), produces higher quality and more grounded responses across various financial domains that are core to Vanguard's business while simultaneously maintaining low costs.
{"title":"MoA is All You Need: Building LLM Research Team using Mixture of Agents","authors":"Sandy Chen, Leqi Zeng, Abhinav Raghunathan, Flora Huang, Terrence C. Kim","doi":"arxiv-2409.07487","DOIUrl":"https://doi.org/arxiv-2409.07487","url":null,"abstract":"Large Language Models (LLMs) research in the financial domain is particularly\u0000complex due to the sheer number of approaches proposed in literature.\u0000Retrieval-Augmented Generation (RAG) has emerged as one of the leading methods\u0000in the sector due to its inherent groundedness and data source variability. In\u0000this work, we introduce a RAG framework called Mixture of Agents (MoA) and\u0000demonstrate its viability as a practical, customizable, and highly effective\u0000approach for scaling RAG applications. MoA is essentially a layered network of\u0000individually customized small language models (Hoffmann et al., 2022)\u0000collaborating to answer questions and extract information. While there are many\u0000theoretical propositions for such an architecture and even a few libraries for\u0000generally applying the structure in practice, there are limited documented\u0000studies evaluating the potential of this framework considering real business\u0000constraints such as cost and speed. We find that the MoA framework, consisting\u0000of small language models (Hoffmann et al., 2022), produces higher quality and\u0000more grounded responses across various financial domains that are core to\u0000Vanguard's business while simultaneously maintaining low costs.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"27 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208125","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}
Managing high-frequency data in a limit order book (LOB) is a complex task that often exceeds the capabilities of conventional time-series forecasting models. Accurately predicting the entire multi-level LOB, beyond just the mid-price, is essential for understanding high-frequency market dynamics. However, this task is challenging due to the complex interdependencies among compound attributes within each dimension, such as order types, features, and levels. In this study, we explore advanced multidimensional sequence-to-sequence models to forecast the entire multi-level LOB, including order prices and volumes. Our main contribution is the development of a compound multivariate embedding method designed to capture the complex relationships between spatiotemporal features. Empirical results show that our method outperforms other multivariate forecasting methods, achieving the lowest forecasting error while preserving the ordinal structure of the LOB.
{"title":"Attention-Based Reading, Highlighting, and Forecasting of the Limit Order Book","authors":"Jiwon Jung, Kiseop Lee","doi":"arxiv-2409.02277","DOIUrl":"https://doi.org/arxiv-2409.02277","url":null,"abstract":"Managing high-frequency data in a limit order book (LOB) is a complex task\u0000that often exceeds the capabilities of conventional time-series forecasting\u0000models. Accurately predicting the entire multi-level LOB, beyond just the\u0000mid-price, is essential for understanding high-frequency market dynamics.\u0000However, this task is challenging due to the complex interdependencies among\u0000compound attributes within each dimension, such as order types, features, and\u0000levels. In this study, we explore advanced multidimensional\u0000sequence-to-sequence models to forecast the entire multi-level LOB, including\u0000order prices and volumes. Our main contribution is the development of a\u0000compound multivariate embedding method designed to capture the complex\u0000relationships between spatiotemporal features. Empirical results show that our\u0000method outperforms other multivariate forecasting methods, achieving the lowest\u0000forecasting error while preserving the ordinal structure of the LOB.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"16 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208123","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}
In the rapidly evolving field of financial forecasting, the application of neural networks presents a compelling advancement over traditional statistical models. This research paper explores the effectiveness of two specific neural forecasting models, N-HiTS and N-BEATS, in predicting financial market trends. Through a systematic comparison with conventional models, this study demonstrates the superior predictive capabilities of neural approaches, particularly in handling the non-linear dynamics and complex patterns inherent in financial time series data. The results indicate that N-HiTS and N-BEATS not only enhance the accuracy of forecasts but also boost the robustness and adaptability of financial predictions, offering substantial advantages in environments that require real-time decision-making. The paper concludes with insights into the practical implications of neural forecasting in financial markets and recommendations for future research directions.
{"title":"Advancing Financial Forecasting: A Comparative Analysis of Neural Forecasting Models N-HiTS and N-BEATS","authors":"Mohit Apte, Yashodhara Haribhakta","doi":"arxiv-2409.00480","DOIUrl":"https://doi.org/arxiv-2409.00480","url":null,"abstract":"In the rapidly evolving field of financial forecasting, the application of\u0000neural networks presents a compelling advancement over traditional statistical\u0000models. This research paper explores the effectiveness of two specific neural\u0000forecasting models, N-HiTS and N-BEATS, in predicting financial market trends.\u0000Through a systematic comparison with conventional models, this study\u0000demonstrates the superior predictive capabilities of neural approaches,\u0000particularly in handling the non-linear dynamics and complex patterns inherent\u0000in financial time series data. The results indicate that N-HiTS and N-BEATS not\u0000only enhance the accuracy of forecasts but also boost the robustness and\u0000adaptability of financial predictions, offering substantial advantages in\u0000environments that require real-time decision-making. The paper concludes with\u0000insights into the practical implications of neural forecasting in financial\u0000markets and recommendations for future research directions.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142208124","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}
This study investigates the application of causal discovery algorithms in equity markets, with a focus on their potential to enhance investment strategies. An investment strategy was developed based on the causal structures identified by these algorithms, and its performance was evaluated to assess the profitability and effectiveness in stock market environments. The results indicate that causal discovery algorithms can successfully uncover actionable causal relationships in large markets, leading to profitable investment outcomes. However, the research also identifies a critical challenge: the computational complexity and scalability of these algorithms when dealing with large datasets, which presents practical limitations for their application in real-world market analysis.
{"title":"Trading with Time Series Causal Discovery: An Empirical Study","authors":"Ruijie Tang","doi":"arxiv-2408.15846","DOIUrl":"https://doi.org/arxiv-2408.15846","url":null,"abstract":"This study investigates the application of causal discovery algorithms in\u0000equity markets, with a focus on their potential to enhance investment\u0000strategies. An investment strategy was developed based on the causal structures\u0000identified by these algorithms, and its performance was evaluated to assess the\u0000profitability and effectiveness in stock market environments. The results\u0000indicate that causal discovery algorithms can successfully uncover actionable\u0000causal relationships in large markets, leading to profitable investment\u0000outcomes. However, the research also identifies a critical challenge: the\u0000computational complexity and scalability of these algorithms when dealing with\u0000large datasets, which presents practical limitations for their application in\u0000real-world market analysis.","PeriodicalId":501294,"journal":{"name":"arXiv - QuantFin - Computational Finance","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142226613","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}