Our research presents a new approach for forecasting the synchronization of stock prices using machine learning and non-linear time-series analysis. To capture the complex non-linear relationships between stock prices, we utilize recurrence plots (RP) and cross-recurrence quantification analysis (CRQA). By transforming Cross Recurrence Plot (CRP) data into a time-series format, we enable the use of Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) networks for predicting stock price synchronization through both regression and classification. We apply this methodology to a dataset of 20 highly capitalized stocks from the Indian market over a 21-year period. The findings reveal that our approach can predict stock price synchronization, with an accuracy of 0.98 and F1 score of 0.83 offering valuable insights for developing effective trading strategies and risk management tools.
{"title":"Leveraging RNNs and LSTMs for Synchronization Analysis in the Indian Stock Market: A Threshold-Based Classification Approach","authors":"Sanjay Sathish, Charu C Sharma","doi":"arxiv-2409.06728","DOIUrl":"https://doi.org/arxiv-2409.06728","url":null,"abstract":"Our research presents a new approach for forecasting the synchronization of\u0000stock prices using machine learning and non-linear time-series analysis. To\u0000capture the complex non-linear relationships between stock prices, we utilize\u0000recurrence plots (RP) and cross-recurrence quantification analysis (CRQA). By\u0000transforming Cross Recurrence Plot (CRP) data into a time-series format, we\u0000enable the use of Recurrent Neural Networks (RNN) and Long Short-Term Memory\u0000(LSTM) networks for predicting stock price synchronization through both\u0000regression and classification. We apply this methodology to a dataset of 20\u0000highly capitalized stocks from the Indian market over a 21-year period. The\u0000findings reveal that our approach can predict stock price synchronization, with\u0000an accuracy of 0.98 and F1 score of 0.83 offering valuable insights for\u0000developing effective trading strategies and risk management tools.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190942","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}
Stock price prediction is a challenging problem in the field of finance and receives widespread attention. In recent years, with the rapid development of technologies such as deep learning and graph neural networks, more research methods have begun to focus on exploring the interrelationships between stocks. However, existing methods mostly focus on the short-term dynamic relationships of stocks and directly integrating relationship information with temporal information. They often overlook the complex nonlinear dynamic characteristics and potential higher-order interaction relationships among stocks in the stock market. Therefore, we propose a stock price trend prediction model named LSR-IGRU in this paper, which is based on long short-term stock relationships and an improved GRU input. Firstly, we construct a long short-term relationship matrix between stocks, where secondary industry information is employed for the first time to capture long-term relationships of stocks, and overnight price information is utilized to establish short-term relationships. Next, we improve the inputs of the GRU model at each step, enabling the model to more effectively integrate temporal information and long short-term relationship information, thereby significantly improving the accuracy of predicting stock trend changes. Finally, through extensive experiments on multiple datasets from stock markets in China and the United States, we validate the superiority of the proposed LSR-IGRU model over the current state-of-the-art baseline models. We also apply the proposed model to the algorithmic trading system of a financial company, achieving significantly higher cumulative portfolio returns compared to other baseline methods. Our sources are released at https://github.com/ZP1481616577/Baselines_LSR-IGRU.
{"title":"LSR-IGRU: Stock Trend Prediction Based on Long Short-Term Relationships and Improved GRU","authors":"Peng Zhu, Yuante Li, Yifan Hu, Qinyuan Liu, Dawei Cheng, Yuqi Liang","doi":"arxiv-2409.08282","DOIUrl":"https://doi.org/arxiv-2409.08282","url":null,"abstract":"Stock price prediction is a challenging problem in the field of finance and\u0000receives widespread attention. In recent years, with the rapid development of\u0000technologies such as deep learning and graph neural networks, more research\u0000methods have begun to focus on exploring the interrelationships between stocks.\u0000However, existing methods mostly focus on the short-term dynamic relationships\u0000of stocks and directly integrating relationship information with temporal\u0000information. They often overlook the complex nonlinear dynamic characteristics\u0000and potential higher-order interaction relationships among stocks in the stock\u0000market. Therefore, we propose a stock price trend prediction model named\u0000LSR-IGRU in this paper, which is based on long short-term stock relationships\u0000and an improved GRU input. Firstly, we construct a long short-term relationship\u0000matrix between stocks, where secondary industry information is employed for the\u0000first time to capture long-term relationships of stocks, and overnight price\u0000information is utilized to establish short-term relationships. Next, we improve\u0000the inputs of the GRU model at each step, enabling the model to more\u0000effectively integrate temporal information and long short-term relationship\u0000information, thereby significantly improving the accuracy of predicting stock\u0000trend changes. Finally, through extensive experiments on multiple datasets from\u0000stock markets in China and the United States, we validate the superiority of\u0000the proposed LSR-IGRU model over the current state-of-the-art baseline models.\u0000We also apply the proposed model to the algorithmic trading system of a\u0000financial company, achieving significantly higher cumulative portfolio returns\u0000compared to other baseline methods. Our sources are released at\u0000https://github.com/ZP1481616577/Baselines_LSR-IGRU.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"15 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142250091","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}
Shengkun Wang, Taoran Ji, Linhan Wang, Yanshen Sun, Shang-Ching Liu, Amit Kumar, Chang-Tien Lu
The stock price prediction task holds a significant role in the financial domain and has been studied for a long time. Recently, large language models (LLMs) have brought new ways to improve these predictions. While recent financial large language models (FinLLMs) have shown considerable progress in financial NLP tasks compared to smaller pre-trained language models (PLMs), challenges persist in stock price forecasting. Firstly, effectively integrating the modalities of time series data and natural language to fully leverage these capabilities remains complex. Secondly, FinLLMs focus more on analysis and interpretability, which can overlook the essential features of time series data. Moreover, due to the abundance of false and redundant information in financial markets, models often produce less accurate predictions when faced with such input data. In this paper, we introduce StockTime, a novel LLM-based architecture designed specifically for stock price data. Unlike recent FinLLMs, StockTime is specifically designed for stock price time series data. It leverages the natural ability of LLMs to predict the next token by treating stock prices as consecutive tokens, extracting textual information such as stock correlations, statistical trends and timestamps directly from these stock prices. StockTime then integrates both textual and time series data into the embedding space. By fusing this multimodal data, StockTime effectively predicts stock prices across arbitrary look-back periods. Our experiments demonstrate that StockTime outperforms recent LLMs, as it gives more accurate predictions while reducing memory usage and runtime costs.
{"title":"StockTime: A Time Series Specialized Large Language Model Architecture for Stock Price Prediction","authors":"Shengkun Wang, Taoran Ji, Linhan Wang, Yanshen Sun, Shang-Ching Liu, Amit Kumar, Chang-Tien Lu","doi":"arxiv-2409.08281","DOIUrl":"https://doi.org/arxiv-2409.08281","url":null,"abstract":"The stock price prediction task holds a significant role in the financial\u0000domain and has been studied for a long time. Recently, large language models\u0000(LLMs) have brought new ways to improve these predictions. While recent\u0000financial large language models (FinLLMs) have shown considerable progress in\u0000financial NLP tasks compared to smaller pre-trained language models (PLMs),\u0000challenges persist in stock price forecasting. Firstly, effectively integrating\u0000the modalities of time series data and natural language to fully leverage these\u0000capabilities remains complex. Secondly, FinLLMs focus more on analysis and\u0000interpretability, which can overlook the essential features of time series\u0000data. Moreover, due to the abundance of false and redundant information in\u0000financial markets, models often produce less accurate predictions when faced\u0000with such input data. In this paper, we introduce StockTime, a novel LLM-based\u0000architecture designed specifically for stock price data. Unlike recent FinLLMs,\u0000StockTime is specifically designed for stock price time series data. It\u0000leverages the natural ability of LLMs to predict the next token by treating\u0000stock prices as consecutive tokens, extracting textual information such as\u0000stock correlations, statistical trends and timestamps directly from these stock\u0000prices. StockTime then integrates both textual and time series data into the\u0000embedding space. By fusing this multimodal data, StockTime effectively predicts\u0000stock prices across arbitrary look-back periods. Our experiments demonstrate\u0000that StockTime outperforms recent LLMs, as it gives more accurate predictions\u0000while reducing memory usage and runtime costs.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"17 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142268344","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}
Baptiste Lefort, Eric Benhamou, Jean-Jacques Ohana, David Saltiel, Beatrice Guez
In this paper, we demonstrate that non-generative, small-sized models such as FinBERT and FinDRoBERTa, when fine-tuned, can outperform GPT-3.5 and GPT-4 models in zero-shot learning settings in sentiment analysis for financial news. These fine-tuned models show comparable results to GPT-3.5 when it is fine-tuned on the task of determining market sentiment from daily financial news summaries sourced from Bloomberg. To fine-tune and compare these models, we created a novel database, which assigns a market score to each piece of news without human interpretation bias, systematically identifying the mentioned companies and analyzing whether their stocks have gone up, down, or remained neutral. Furthermore, the paper shows that the assumptions of Condorcet's Jury Theorem do not hold suggesting that fine-tuned small models are not independent of the fine-tuned GPT models, indicating behavioural similarities. Lastly, the resulted fine-tuned models are made publicly available on HuggingFace, providing a resource for further research in financial sentiment analysis and text classification.
{"title":"Optimizing Performance: How Compact Models Match or Exceed GPT's Classification Capabilities through Fine-Tuning","authors":"Baptiste Lefort, Eric Benhamou, Jean-Jacques Ohana, David Saltiel, Beatrice Guez","doi":"arxiv-2409.11408","DOIUrl":"https://doi.org/arxiv-2409.11408","url":null,"abstract":"In this paper, we demonstrate that non-generative, small-sized models such as\u0000FinBERT and FinDRoBERTa, when fine-tuned, can outperform GPT-3.5 and GPT-4\u0000models in zero-shot learning settings in sentiment analysis for financial news.\u0000These fine-tuned models show comparable results to GPT-3.5 when it is\u0000fine-tuned on the task of determining market sentiment from daily financial\u0000news summaries sourced from Bloomberg. To fine-tune and compare these models,\u0000we created a novel database, which assigns a market score to each piece of news\u0000without human interpretation bias, systematically identifying the mentioned\u0000companies and analyzing whether their stocks have gone up, down, or remained\u0000neutral. Furthermore, the paper shows that the assumptions of Condorcet's Jury\u0000Theorem do not hold suggesting that fine-tuned small models are not independent\u0000of the fine-tuned GPT models, indicating behavioural similarities. Lastly, the\u0000resulted fine-tuned models are made publicly available on HuggingFace,\u0000providing a resource for further research in financial sentiment analysis and\u0000text classification.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"17 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142250090","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}
Recent advancements in Distributional Reinforcement Learning (DRL) for modeling loss distributions have shown promise in developing hedging strategies in derivatives markets. A common approach in DRL involves learning the quantiles of loss distributions at specified levels using Quantile Regression (QR). This method is particularly effective in option hedging due to its direct quantile-based risk assessment, such as Value at Risk (VaR) and Conditional Value at Risk (CVaR). However, these risk measures depend on the accurate estimation of extreme quantiles in the loss distribution's tail, which can be imprecise in QR-based DRL due to the rarity and extremity of tail data, as highlighted in the literature. To address this issue, we propose EXtreme DRL (EX-DRL), which enhances extreme quantile prediction by modeling the tail of the loss distribution with a Generalized Pareto Distribution (GPD). This method introduces supplementary data to mitigate the scarcity of extreme quantile observations, thereby improving estimation accuracy through QR. Comprehensive experiments on gamma hedging options demonstrate that EX-DRL improves existing QR-based models by providing more precise estimates of extreme quantiles, thereby improving the computation and reliability of risk metrics for complex financial risk management.
分布强化学习（DRL）在损失分布建模方面的最新进展，为衍生品市场对冲策略的开发带来了希望。DRL 中的一种常见方法是利用定量回归（QR）学习指定水平上损失分布的定量。这种方法在期权对冲中尤为有效，因为它可以直接进行基于量值的风险评估，如风险值（VaR）和条件风险值（CVaR）。然而，这些风险度量依赖于对损失分布尾部极端量值的精确估计，由于尾部数据的稀缺性和极端性，基于 QR 的 DRL 难以精确估计尾部数据。为了解决这个问题，我们提出了 EXtreme DRL（EX-DRL），它通过使用广义帕累托分布（GPD）对损失分布的尾部进行建模，从而增强了极端量值预测。该方法引入了补充数据，以缓解极端量级观测数据稀缺的问题，从而通过 QR 提高了估计精度。对伽马对冲期权的综合实验表明，EX-DRL 改进了现有的基于 QR 的模型，提供了更精确的极端量值估计，从而改进了用于综合金融风险管理的风险度量的计算和可靠性。
{"title":"EX-DRL: Hedging Against Heavy Losses with EXtreme Distributional Reinforcement Learning","authors":"Parvin Malekzadeh, Zissis Poulos, Jacky Chen, Zeyu Wang, Konstantinos N. Plataniotis","doi":"arxiv-2408.12446","DOIUrl":"https://doi.org/arxiv-2408.12446","url":null,"abstract":"Recent advancements in Distributional Reinforcement Learning (DRL) for\u0000modeling loss distributions have shown promise in developing hedging strategies\u0000in derivatives markets. A common approach in DRL involves learning the\u0000quantiles of loss distributions at specified levels using Quantile Regression\u0000(QR). This method is particularly effective in option hedging due to its direct\u0000quantile-based risk assessment, such as Value at Risk (VaR) and Conditional\u0000Value at Risk (CVaR). However, these risk measures depend on the accurate\u0000estimation of extreme quantiles in the loss distribution's tail, which can be\u0000imprecise in QR-based DRL due to the rarity and extremity of tail data, as\u0000highlighted in the literature. To address this issue, we propose EXtreme DRL\u0000(EX-DRL), which enhances extreme quantile prediction by modeling the tail of\u0000the loss distribution with a Generalized Pareto Distribution (GPD). This method\u0000introduces supplementary data to mitigate the scarcity of extreme quantile\u0000observations, thereby improving estimation accuracy through QR. Comprehensive\u0000experiments on gamma hedging options demonstrate that EX-DRL improves existing\u0000QR-based models by providing more precise estimates of extreme quantiles,\u0000thereby improving the computation and reliability of risk metrics for complex\u0000financial risk management.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"121 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190917","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}
Applying a network analysis to stock return correlations, we study the dynamical properties of the network and how they correlate with the market return, finding meaningful variables that partially capture the complex dynamical processes of stock interactions and the market structure. We then use the individual properties of stocks within the network along with the global ones, to find correlations with the future returns of individual S&P 500 stocks. Applying these properties as input variables for forecasting, we find a 50% improvement on the R2score in the prediction of stock returns on long time scales (per year), and 3% on short time scales (2 days), relative to baseline models without network variables.
{"title":"Dynamical analysis of financial stocks network: improving forecasting using network properties","authors":"Ixandra Achitouv","doi":"arxiv-2408.11759","DOIUrl":"https://doi.org/arxiv-2408.11759","url":null,"abstract":"Applying a network analysis to stock return correlations, we study the\u0000dynamical properties of the network and how they correlate with the market\u0000return, finding meaningful variables that partially capture the complex\u0000dynamical processes of stock interactions and the market structure. We then use\u0000the individual properties of stocks within the network along with the global\u0000ones, to find correlations with the future returns of individual S&P 500\u0000stocks. Applying these properties as input variables for forecasting, we find a\u000050% improvement on the R2score in the prediction of stock returns on long time\u0000scales (per year), and 3% on short time scales (2 days), relative to baseline\u0000models without network variables.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"51 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190915","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}
Gabriel Rodrigues Palma, Mariusz Skoczeń, Phil Maguire
The decisions traders make to buy or sell an asset depend on various analyses, with expertise required to identify patterns that can be exploited for profit. In this paper we identify novel features extracted from emergent and well-established financial markets using linear models and Gaussian Mixture Models (GMM) with the aim of finding profitable opportunities. We used approximately six months of data consisting of minute candles from the Bitcoin, Pepecoin, and Nasdaq markets to derive and compare the proposed novel features with commonly used ones. These features were extracted based on the previous 59 minutes for each market and used to identify predictions for the hour ahead. We explored the performance of various machine learning strategies, such as Random Forests (RF) and K-Nearest Neighbours (KNN) to classify market movements. A naive random approach to selecting trading decisions was used as a benchmark, with outcomes assumed to be equally likely. We used a temporal cross-validation approach using test sets of 40%, 30% and 20% of total hours to evaluate the learning algorithms' performances. Our results showed that filtering the time series facilitates algorithms' generalisation. The GMM filtering approach revealed that the KNN and RF algorithms produced higher average returns than the random algorithm.
{"title":"Combining supervised and unsupervised learning methods to predict financial market movements","authors":"Gabriel Rodrigues Palma, Mariusz Skoczeń, Phil Maguire","doi":"arxiv-2409.03762","DOIUrl":"https://doi.org/arxiv-2409.03762","url":null,"abstract":"The decisions traders make to buy or sell an asset depend on various\u0000analyses, with expertise required to identify patterns that can be exploited\u0000for profit. In this paper we identify novel features extracted from emergent\u0000and well-established financial markets using linear models and Gaussian Mixture\u0000Models (GMM) with the aim of finding profitable opportunities. We used\u0000approximately six months of data consisting of minute candles from the Bitcoin,\u0000Pepecoin, and Nasdaq markets to derive and compare the proposed novel features\u0000with commonly used ones. These features were extracted based on the previous 59\u0000minutes for each market and used to identify predictions for the hour ahead. We\u0000explored the performance of various machine learning strategies, such as Random\u0000Forests (RF) and K-Nearest Neighbours (KNN) to classify market movements. A\u0000naive random approach to selecting trading decisions was used as a benchmark,\u0000with outcomes assumed to be equally likely. We used a temporal cross-validation\u0000approach using test sets of 40%, 30% and 20% of total hours to evaluate the\u0000learning algorithms' performances. Our results showed that filtering the time\u0000series facilitates algorithms' generalisation. The GMM filtering approach\u0000revealed that the KNN and RF algorithms produced higher average returns than\u0000the random algorithm.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"9 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190940","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}
We use Fourier analysis to access risk in financial products. With it we analyze price changes of e.g. stocks. Via Fourier analysis we scrutinize quantitatively whether the frequency of change is higher than a change in (conserved) company value would allow. If it is the case, it would be a clear indicator of speculation and with it risk. The entire methods or better its application is fairly new. However, there were severe flaws in previous attempts; making the results (not the method) doubtful. We corrected all these mistakes by e.g. using Fourier transformation instead of discrete Fourier analysis. Our analysis is reliable in the entire frequency band, even for fre-quency of 1/1d or higher if the prices are noted accordingly. For the stocks scrutinized we found that the price of stocks changes disproportionally within one week which clearly indicates spec-ulation. It would be an interesting extension to apply the method to crypto currencies as these currencies have no conserved value which makes normal considerations of volatility difficult.
{"title":"A new measure of risk using Fourier analysis","authors":"Michael Grabinski, Galiya Klinkova","doi":"arxiv-2408.10279","DOIUrl":"https://doi.org/arxiv-2408.10279","url":null,"abstract":"We use Fourier analysis to access risk in financial products. With it we\u0000analyze price changes of e.g. stocks. Via Fourier analysis we scrutinize\u0000quantitatively whether the frequency of change is higher than a change in\u0000(conserved) company value would allow. If it is the case, it would be a clear\u0000indicator of speculation and with it risk. The entire methods or better its\u0000application is fairly new. However, there were severe flaws in previous\u0000attempts; making the results (not the method) doubtful. We corrected all these\u0000mistakes by e.g. using Fourier transformation instead of discrete Fourier\u0000analysis. Our analysis is reliable in the entire frequency band, even for\u0000fre-quency of 1/1d or higher if the prices are noted accordingly. For the\u0000stocks scrutinized we found that the price of stocks changes disproportionally\u0000within one week which clearly indicates spec-ulation. It would be an\u0000interesting extension to apply the method to crypto currencies as these\u0000currencies have no conserved value which makes normal considerations of\u0000volatility difficult.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"7 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190916","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 paper presents a comprehensive statistical analysis of the Web3 ecosystem, comparing various Web3 tokens with traditional financial assets across multiple time scales. We examine probability distributions, tail behaviors, and other key stylized facts of the returns for a diverse range of tokens, including decentralized exchanges, liquidity pools, and centralized exchanges. Despite functional differences, most tokens exhibit well-established empirical facts, including unconditional probability density of returns with heavy tails gradually becoming Gaussian and volatility clustering. Furthermore, we compare assets traded on centralized (CEX) and decentralized (DEX) exchanges, finding that DEXs exhibit similar stylized facts despite different trading mechanisms and often divergent long-term performance. We propose that this similarity is attributable to arbitrageurs striving to maintain similar centralized and decentralized prices. Our study contributes to a better understanding of the dynamics of Web3 tokens and the relationship between CEX and DEX markets, with important implications for risk management, pricing models, and portfolio construction in the rapidly evolving DeFi landscape. These results add to the growing body of literature on cryptocurrency markets and provide insights that can guide the development of more accurate models for DeFi markets.
{"title":"Stylized facts in Web3","authors":"A. Christian Silva, Shen-Ning Tung, Wwi-Ru Chen","doi":"arxiv-2408.07653","DOIUrl":"https://doi.org/arxiv-2408.07653","url":null,"abstract":"This paper presents a comprehensive statistical analysis of the Web3\u0000ecosystem, comparing various Web3 tokens with traditional financial assets\u0000across multiple time scales. We examine probability distributions, tail\u0000behaviors, and other key stylized facts of the returns for a diverse range of\u0000tokens, including decentralized exchanges, liquidity pools, and centralized\u0000exchanges. Despite functional differences, most tokens exhibit well-established\u0000empirical facts, including unconditional probability density of returns with\u0000heavy tails gradually becoming Gaussian and volatility clustering. Furthermore,\u0000we compare assets traded on centralized (CEX) and decentralized (DEX)\u0000exchanges, finding that DEXs exhibit similar stylized facts despite different\u0000trading mechanisms and often divergent long-term performance. We propose that\u0000this similarity is attributable to arbitrageurs striving to maintain similar\u0000centralized and decentralized prices. Our study contributes to a better\u0000understanding of the dynamics of Web3 tokens and the relationship between CEX\u0000and DEX markets, with important implications for risk management, pricing\u0000models, and portfolio construction in the rapidly evolving DeFi landscape.\u0000These results add to the growing body of literature on cryptocurrency markets\u0000and provide insights that can guide the development of more accurate models for\u0000DeFi markets.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"33 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190935","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 objective of this work is the investigation of complexity, asymmetry, stochasticity and non-linearity of the financial and economic systems by using the tools of statistical mechanics and information theory. More precisely, this thesis concerns statistical-based modeling and empirical analyses with applications in finance, forecasting, production processes and game theory. In these areas the time dependence of probability distributions is of prime interest and can be measured or exactly calculated for model systems. The correlation coefficients and moments are among the useful quantities to describe the dynamics and the correlations between random variables. However, the full investigation can only be achieved if the probability distribution function of the variable is known; its derivation is one of the main focuses of the present work.
{"title":"Model-based and empirical analyses of stochastic fluctuations in economy and finance","authors":"Rubina Zadourian","doi":"arxiv-2408.16010","DOIUrl":"https://doi.org/arxiv-2408.16010","url":null,"abstract":"The objective of this work is the investigation of complexity, asymmetry,\u0000stochasticity and non-linearity of the financial and economic systems by using\u0000the tools of statistical mechanics and information theory. More precisely, this\u0000thesis concerns statistical-based modeling and empirical analyses with\u0000applications in finance, forecasting, production processes and game theory. In\u0000these areas the time dependence of probability distributions is of prime\u0000interest and can be measured or exactly calculated for model systems. The\u0000correlation coefficients and moments are among the useful quantities to\u0000describe the dynamics and the correlations between random variables. However,\u0000the full investigation can only be achieved if the probability distribution\u0000function of the variable is known; its derivation is one of the main focuses of\u0000the present work.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142190938","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：