Jiajian Zheng, Duan Xin, Qishuo Cheng, Miao Tian, Le Yang
The stock market is a crucial component of the financial market, playing a vital role in wealth accumulation for investors, financing costs for listed companies, and the stable development of the national macroeconomy. Significant fluctuations in the stock market can damage the interests of stock investors and cause an imbalance in the industrial structure, which can interfere with the macro level development of the national economy. The prediction of stock price trends is a popular research topic in academia. Predicting the three trends of stock pricesrising, sideways, and falling can assist investors in making informed decisions about buying, holding, or selling stocks. Establishing an effective forecasting model for predicting these trends is of substantial practical importance. This paper evaluates the predictive performance of random forest models combined with artificial intelligence on a test set of four stocks using optimal parameters. The evaluation considers both predictive accuracy and time efficiency.
{"title":"The Random Forest Model for Analyzing and Forecasting the US Stock Market in the Context of Smart Finance","authors":"Jiajian Zheng, Duan Xin, Qishuo Cheng, Miao Tian, Le Yang","doi":"arxiv-2402.17194","DOIUrl":"https://doi.org/arxiv-2402.17194","url":null,"abstract":"The stock market is a crucial component of the financial market, playing a\u0000vital role in wealth accumulation for investors, financing costs for listed\u0000companies, and the stable development of the national macroeconomy. Significant\u0000fluctuations in the stock market can damage the interests of stock investors\u0000and cause an imbalance in the industrial structure, which can interfere with\u0000the macro level development of the national economy. The prediction of stock\u0000price trends is a popular research topic in academia. Predicting the three\u0000trends of stock pricesrising, sideways, and falling can assist investors in\u0000making informed decisions about buying, holding, or selling stocks.\u0000Establishing an effective forecasting model for predicting these trends is of\u0000substantial practical importance. This paper evaluates the predictive\u0000performance of random forest models combined with artificial intelligence on a\u0000test set of four stocks using optimal parameters. The evaluation considers both\u0000predictive accuracy and time efficiency.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"13 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140009765","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 delves into the analysis of financial markets through the lens of Wyckoff Phases, a framework devised by Richard D. Wyckoff in the early 20th century. Focusing on the accumulation pattern within the Wyckoff framework, the research explores the phases of trading range and secondary test, elucidating their significance in understanding market dynamics and identifying potential trading opportunities. By dissecting the intricacies of these phases, the study sheds light on the creation of liquidity through market structure, offering insights into how traders can leverage this knowledge to anticipate price movements and make informed decisions. The effective detection and analysis of Wyckoff patterns necessitate robust computational models capable of processing complex market data, with spatial data best analyzed using Convolutional Neural Networks (CNNs) and temporal data through Long Short-Term Memory (LSTM) models. The creation of training data involves the generation of swing points, representing significant market movements, and filler points, introducing noise and enhancing model generalization. Activation functions, such as the sigmoid function, play a crucial role in determining the output behavior of neural network models. The results of the study demonstrate the remarkable efficacy of deep learning models in detecting Wyckoff patterns within financial data, underscoring their potential for enhancing pattern recognition and analysis in financial markets. In conclusion, the study highlights the transformative potential of AI-driven approaches in financial analysis and trading strategies, with the integration of AI technologies shaping the future of trading and investment practices.
{"title":"Long Short-Term Memory Pattern Recognition in Currency Trading","authors":"Jai Pal","doi":"arxiv-2403.18839","DOIUrl":"https://doi.org/arxiv-2403.18839","url":null,"abstract":"This study delves into the analysis of financial markets through the lens of\u0000Wyckoff Phases, a framework devised by Richard D. Wyckoff in the early 20th\u0000century. Focusing on the accumulation pattern within the Wyckoff framework, the\u0000research explores the phases of trading range and secondary test, elucidating\u0000their significance in understanding market dynamics and identifying potential\u0000trading opportunities. By dissecting the intricacies of these phases, the study\u0000sheds light on the creation of liquidity through market structure, offering\u0000insights into how traders can leverage this knowledge to anticipate price\u0000movements and make informed decisions. The effective detection and analysis of\u0000Wyckoff patterns necessitate robust computational models capable of processing\u0000complex market data, with spatial data best analyzed using Convolutional Neural\u0000Networks (CNNs) and temporal data through Long Short-Term Memory (LSTM) models.\u0000The creation of training data involves the generation of swing points,\u0000representing significant market movements, and filler points, introducing noise\u0000and enhancing model generalization. Activation functions, such as the sigmoid\u0000function, play a crucial role in determining the output behavior of neural\u0000network models. The results of the study demonstrate the remarkable efficacy of\u0000deep learning models in detecting Wyckoff patterns within financial data,\u0000underscoring their potential for enhancing pattern recognition and analysis in\u0000financial markets. In conclusion, the study highlights the transformative\u0000potential of AI-driven approaches in financial analysis and trading strategies,\u0000with the integration of AI technologies shaping the future of trading and\u0000investment practices.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"82 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140323436","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}
Optimal execution is an important problem faced by any trader. Most solutions are based on the assumption of constant market impact, while liquidity is known to be dynamic. Moreover, models with time-varying liquidity typically assume that it is observable, despite the fact that, in reality, it is latent and hard to measure in real time. In this paper we show that the use of Double Deep Q-learning, a form of Reinforcement Learning based on neural networks, is able to learn optimal trading policies when liquidity is time-varying. Specifically, we consider an Almgren-Chriss framework with temporary and permanent impact parameters following several deterministic and stochastic dynamics. Using extensive numerical experiments, we show that the trained algorithm learns the optimal policy when the analytical solution is available, and overcomes benchmarks and approximated solutions when the solution is not available.
{"title":"Reinforcement Learning for Optimal Execution when Liquidity is Time-Varying","authors":"Andrea Macrì, Fabrizio Lillo","doi":"arxiv-2402.12049","DOIUrl":"https://doi.org/arxiv-2402.12049","url":null,"abstract":"Optimal execution is an important problem faced by any trader. Most solutions\u0000are based on the assumption of constant market impact, while liquidity is known\u0000to be dynamic. Moreover, models with time-varying liquidity typically assume\u0000that it is observable, despite the fact that, in reality, it is latent and hard\u0000to measure in real time. In this paper we show that the use of Double Deep\u0000Q-learning, a form of Reinforcement Learning based on neural networks, is able\u0000to learn optimal trading policies when liquidity is time-varying. Specifically,\u0000we consider an Almgren-Chriss framework with temporary and permanent impact\u0000parameters following several deterministic and stochastic dynamics. Using\u0000extensive numerical experiments, we show that the trained algorithm learns the\u0000optimal policy when the analytical solution is available, and overcomes\u0000benchmarks and approximated solutions when the solution is not available.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"35 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139928078","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}
Convex optimisation has provided a mechanism to determine arbitrage trades on automated market markets (AMMs) since almost their inception. Here we outline generic closed-form solutions for $N$-token geometric mean market maker pool arbitrage, that in simulation (with synthetic and historic data) provide better arbitrage opportunities than convex optimisers and is able to capitalise on those opportunities sooner. Furthermore, the intrinsic parallelism of the proposed approach (unlike convex optimisation) offers the ability to scale on GPUs, opening up a new approach to AMM modelling by offering an alternative to numerical-solver-based methods. The lower computational cost of running this new mechanism can also enable on-chain arbitrage bots for multi-asset pools.
{"title":"Closed-form solutions for generic N-token AMM arbitrage","authors":"Matthew Willetts, Christian Harrington","doi":"arxiv-2402.06731","DOIUrl":"https://doi.org/arxiv-2402.06731","url":null,"abstract":"Convex optimisation has provided a mechanism to determine arbitrage trades on\u0000automated market markets (AMMs) since almost their inception. Here we outline\u0000generic closed-form solutions for $N$-token geometric mean market maker pool\u0000arbitrage, that in simulation (with synthetic and historic data) provide better\u0000arbitrage opportunities than convex optimisers and is able to capitalise on\u0000those opportunities sooner. Furthermore, the intrinsic parallelism of the\u0000proposed approach (unlike convex optimisation) offers the ability to scale on\u0000GPUs, opening up a new approach to AMM modelling by offering an alternative to\u0000numerical-solver-based methods. The lower computational cost of running this\u0000new mechanism can also enable on-chain arbitrage bots for multi-asset pools.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"7 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139760680","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 this paper, we introduce DeepTraderX (DTX), a simple Deep Learning-based trader, and present results that demonstrate its performance in a multi-threaded market simulation. In a total of about 500 simulated market days, DTX has learned solely by watching the prices that other strategies produce. By doing this, it has successfully created a mapping from market data to quotes, either bid or ask orders, to place for an asset. Trained on historical Level-2 market data, i.e., the Limit Order Book (LOB) for specific tradable assets, DTX processes the market state $S$ at each timestep $T$ to determine a price $P$ for market orders. The market data used in both training and testing was generated from unique market schedules based on real historic stock market data. DTX was tested extensively against the best strategies in the literature, with its results validated by statistical analysis. Our findings underscore DTX's capability to rival, and in many instances, surpass, the performance of public-domain traders, including those that outclass human traders, emphasising the efficiency of simple models, as this is required to succeed in intricate multi-threaded simulations. This highlights the potential of leveraging "black-box" Deep Learning systems to create more efficient financial markets.
{"title":"DeepTraderX: Challenging Conventional Trading Strategies with Deep Learning in Multi-Threaded Market Simulations","authors":"Armand Mihai Cismaru","doi":"arxiv-2403.18831","DOIUrl":"https://doi.org/arxiv-2403.18831","url":null,"abstract":"In this paper, we introduce DeepTraderX (DTX), a simple Deep Learning-based\u0000trader, and present results that demonstrate its performance in a\u0000multi-threaded market simulation. In a total of about 500 simulated market\u0000days, DTX has learned solely by watching the prices that other strategies\u0000produce. By doing this, it has successfully created a mapping from market data\u0000to quotes, either bid or ask orders, to place for an asset. Trained on\u0000historical Level-2 market data, i.e., the Limit Order Book (LOB) for specific\u0000tradable assets, DTX processes the market state $S$ at each timestep $T$ to\u0000determine a price $P$ for market orders. The market data used in both training\u0000and testing was generated from unique market schedules based on real historic\u0000stock market data. DTX was tested extensively against the best strategies in\u0000the literature, with its results validated by statistical analysis. Our\u0000findings underscore DTX's capability to rival, and in many instances, surpass,\u0000the performance of public-domain traders, including those that outclass human\u0000traders, emphasising the efficiency of simple models, as this is required to\u0000succeed in intricate multi-threaded simulations. This highlights the potential\u0000of leveraging \"black-box\" Deep Learning systems to create more efficient\u0000financial markets.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"34 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140323284","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 presents a groundbreaking Systematization of Knowledge (SoK) initiative, focusing on an in-depth exploration of the dynamics and behavior of traders on perpetual future contracts across both centralized exchanges (CEXs), and decentralized exchanges (DEXs). We have refined the existing model for investigating traders' behavior in reaction to price volatility to create a new analytical framework specifically for these contract platforms, while also highlighting the role of blockchain technology in their application. Our research includes a comparative analysis of historical data from CEXs and a more extensive examination of complete transactional data on DEXs. On DEX of Virtual Automated Market Making (VAMM) Model, open interest on short and long positions exert effect on price volatility in opposite direction, attributable to VAMM's price formation mechanism. In the DEXs with Oracle Pricing Model, we observed a distinct asymmetry in trader behavior between buyers and sellers. Such asymmetry might stem from uninformed traders reacting more strongly to positive news than to negative, leading to a tendency to accumulate long positions. This study sheds light on the potential risks and advantages of using perpetual future contracts within the DeFi space while provides mathematical basis and empirical insights based on which future theoretical works can be configurated, offering crucial insights into the rapidly evolving world of blockchain-based financial instruments.
{"title":"Perpetual Future Contracts in Centralized and Decentralized Exchanges: Mechanism and Traders' Behavior","authors":"Erdong Chen, Mengzhong Ma, Zixin Nie","doi":"arxiv-2402.03953","DOIUrl":"https://doi.org/arxiv-2402.03953","url":null,"abstract":"This study presents a groundbreaking Systematization of Knowledge (SoK)\u0000initiative, focusing on an in-depth exploration of the dynamics and behavior of\u0000traders on perpetual future contracts across both centralized exchanges (CEXs),\u0000and decentralized exchanges (DEXs). We have refined the existing model for\u0000investigating traders' behavior in reaction to price volatility to create a new\u0000analytical framework specifically for these contract platforms, while also\u0000highlighting the role of blockchain technology in their application. Our\u0000research includes a comparative analysis of historical data from CEXs and a\u0000more extensive examination of complete transactional data on DEXs. On DEX of\u0000Virtual Automated Market Making (VAMM) Model, open interest on short and long\u0000positions exert effect on price volatility in opposite direction, attributable\u0000to VAMM's price formation mechanism. In the DEXs with Oracle Pricing Model, we\u0000observed a distinct asymmetry in trader behavior between buyers and sellers.\u0000Such asymmetry might stem from uninformed traders reacting more strongly to\u0000positive news than to negative, leading to a tendency to accumulate long\u0000positions. This study sheds light on the potential risks and advantages of\u0000using perpetual future contracts within the DeFi space while provides\u0000mathematical basis and empirical insights based on which future theoretical\u0000works can be configurated, offering crucial insights into the rapidly evolving\u0000world of blockchain-based financial instruments.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"22 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139760763","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}
Andrew Ye, James Xu, Yi Wang, Yifan Yu, Daniel Yan, Ryan Chen, Bosheng Dong, Vipin Chaudhary, Shuai Xu
We propose the integration of sentiment analysis and deep-reinforcement learning ensemble algorithms for stock trading, and design a strategy capable of dynamically altering its employed agent given concurrent market sentiment. In particular, we create a simple-yet-effective method for extracting news sentiment and combine this with general improvements upon existing works, resulting in automated trading agents that effectively consider both qualitative market factors and quantitative stock data. We show that our approach results in a strategy that is profitable, robust, and risk-minimal -- outperforming the traditional ensemble strategy as well as single agent algorithms and market metrics. Our findings determine that the conventional practice of switching ensemble agents every fixed-number of months is sub-optimal, and that a dynamic sentiment-based framework greatly unlocks additional performance within these agents. Furthermore, as we have designed our algorithm with simplicity and efficiency in mind, we hypothesize that the transition of our method from historical evaluation towards real-time trading with live data should be relatively simple.
{"title":"Learning the Market: Sentiment-Based Ensemble Trading Agents","authors":"Andrew Ye, James Xu, Yi Wang, Yifan Yu, Daniel Yan, Ryan Chen, Bosheng Dong, Vipin Chaudhary, Shuai Xu","doi":"arxiv-2402.01441","DOIUrl":"https://doi.org/arxiv-2402.01441","url":null,"abstract":"We propose the integration of sentiment analysis and deep-reinforcement\u0000learning ensemble algorithms for stock trading, and design a strategy capable\u0000of dynamically altering its employed agent given concurrent market sentiment.\u0000In particular, we create a simple-yet-effective method for extracting news\u0000sentiment and combine this with general improvements upon existing works,\u0000resulting in automated trading agents that effectively consider both\u0000qualitative market factors and quantitative stock data. We show that our\u0000approach results in a strategy that is profitable, robust, and risk-minimal --\u0000outperforming the traditional ensemble strategy as well as single agent\u0000algorithms and market metrics. Our findings determine that the conventional\u0000practice of switching ensemble agents every fixed-number of months is\u0000sub-optimal, and that a dynamic sentiment-based framework greatly unlocks\u0000additional performance within these agents. Furthermore, as we have designed\u0000our algorithm with simplicity and efficiency in mind, we hypothesize that the\u0000transition of our method from historical evaluation towards real-time trading\u0000with live data should be relatively simple.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"60 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139690179","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}
Eeshaan Dutta, Sarthak Diwan, Siddhartha P. Chakrabarty
This paper proposes an algorithmic trading framework integrating Environmental, Social, and Governance (ESG) ratings with a pairs trading strategy. It addresses the demand for socially responsible investment solutions by developing a unique algorithm blending ESG data with methods for identifying co-integrated stocks. This allows selecting profitable pairs adhering to ESG principles. Further, it incorporates technical indicators for optimal trade execution within this sustainability framework. Extensive back-testing provides evidence of the model's effectiveness, consistently generating positive returns exceeding conventional pairs trading strategies, while upholding ESG principles. This paves the way for a transformative approach to algorithmic trading, offering insights for investors, policymakers, and academics.
{"title":"ESG driven pairs algorithm for sustainable trading: Analysis from the Indian market","authors":"Eeshaan Dutta, Sarthak Diwan, Siddhartha P. Chakrabarty","doi":"arxiv-2401.14761","DOIUrl":"https://doi.org/arxiv-2401.14761","url":null,"abstract":"This paper proposes an algorithmic trading framework integrating\u0000Environmental, Social, and Governance (ESG) ratings with a pairs trading\u0000strategy. It addresses the demand for socially responsible investment solutions\u0000by developing a unique algorithm blending ESG data with methods for identifying\u0000co-integrated stocks. This allows selecting profitable pairs adhering to ESG\u0000principles. Further, it incorporates technical indicators for optimal trade\u0000execution within this sustainability framework. Extensive back-testing provides\u0000evidence of the model's effectiveness, consistently generating positive returns\u0000exceeding conventional pairs trading strategies, while upholding ESG\u0000principles. This paves the way for a transformative approach to algorithmic\u0000trading, offering insights for investors, policymakers, and academics.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"38 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139580771","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 an in-depth analysis of stylized facts in the context of futures on German bonds. The study examines four futures contracts on German bonds: Schatz, Bobl, Bund and Buxl, using tick-by-tick limit order book datasets. It uncovers a range of stylized facts and empirical observations, including the distribution of order sizes, patterns of order flow, and inter-arrival times of orders. The findings reveal both commonalities and unique characteristics across the different futures, thereby enriching our understanding of these markets. Furthermore, the paper introduces insightful realism metrics that can be used to benchmark market simulators. The study contributes to the literature on financial stylized facts by extending empirical observations to this class of assets, which has been relatively underexplored in existing research. This work provides valuable guidance for the development of more accurate and realistic market simulators.
{"title":"Stylized Facts and Market Microstructure: An In-Depth Exploration of German Bond Futures Market","authors":"Hamza Bodor, Laurent Carlier","doi":"arxiv-2401.10722","DOIUrl":"https://doi.org/arxiv-2401.10722","url":null,"abstract":"This paper presents an in-depth analysis of stylized facts in the context of\u0000futures on German bonds. The study examines four futures contracts on German\u0000bonds: Schatz, Bobl, Bund and Buxl, using tick-by-tick limit order book\u0000datasets. It uncovers a range of stylized facts and empirical observations,\u0000including the distribution of order sizes, patterns of order flow, and\u0000inter-arrival times of orders. The findings reveal both commonalities and\u0000unique characteristics across the different futures, thereby enriching our\u0000understanding of these markets. Furthermore, the paper introduces insightful\u0000realism metrics that can be used to benchmark market simulators. The study\u0000contributes to the literature on financial stylized facts by extending\u0000empirical observations to this class of assets, which has been relatively\u0000underexplored in existing research. This work provides valuable guidance for\u0000the development of more accurate and realistic market simulators.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"14 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139515129","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}
Valentina Aparicio, Daniel Gordon, Sebastian G. Huayamares, Yuhuai Luo
Large language models (LLMs) are deep learning algorithms being used to perform natural language processing tasks in various fields, from social sciences to finance and biomedical sciences. Developing and training a new LLM can be very computationally expensive, so it is becoming a common practice to take existing LLMs and finetune them with carefully curated datasets for desired applications in different fields. Here, we present BioFinBERT, a finetuned LLM to perform financial sentiment analysis of public text associated with stocks of companies in the biotechnology sector. The stocks of biotech companies developing highly innovative and risky therapeutic drugs tend to respond very positively or negatively upon a successful or failed clinical readout or regulatory approval of their drug, respectively. These clinical or regulatory results are disclosed by the biotech companies via press releases, which are followed by a significant stock response in many cases. In our attempt to design a LLM capable of analyzing the sentiment of these press releases,we first finetuned BioBERT, a biomedical language representation model designed for biomedical text mining, using financial textual databases. Our finetuned model, termed BioFinBERT, was then used to perform financial sentiment analysis of various biotech-related press releases and financial text around inflection points that significantly affected the price of biotech stocks.
{"title":"BioFinBERT: Finetuning Large Language Models (LLMs) to Analyze Sentiment of Press Releases and Financial Text Around Inflection Points of Biotech Stocks","authors":"Valentina Aparicio, Daniel Gordon, Sebastian G. Huayamares, Yuhuai Luo","doi":"arxiv-2401.11011","DOIUrl":"https://doi.org/arxiv-2401.11011","url":null,"abstract":"Large language models (LLMs) are deep learning algorithms being used to\u0000perform natural language processing tasks in various fields, from social\u0000sciences to finance and biomedical sciences. Developing and training a new LLM\u0000can be very computationally expensive, so it is becoming a common practice to\u0000take existing LLMs and finetune them with carefully curated datasets for\u0000desired applications in different fields. Here, we present BioFinBERT, a\u0000finetuned LLM to perform financial sentiment analysis of public text associated\u0000with stocks of companies in the biotechnology sector. The stocks of biotech\u0000companies developing highly innovative and risky therapeutic drugs tend to\u0000respond very positively or negatively upon a successful or failed clinical\u0000readout or regulatory approval of their drug, respectively. These clinical or\u0000regulatory results are disclosed by the biotech companies via press releases,\u0000which are followed by a significant stock response in many cases. In our\u0000attempt to design a LLM capable of analyzing the sentiment of these press\u0000releases,we first finetuned BioBERT, a biomedical language representation model\u0000designed for biomedical text mining, using financial textual databases. Our\u0000finetuned model, termed BioFinBERT, was then used to perform financial\u0000sentiment analysis of various biotech-related press releases and financial text\u0000around inflection points that significantly affected the price of biotech\u0000stocks.","PeriodicalId":501478,"journal":{"name":"arXiv - QuantFin - Trading and Market Microstructure","volume":"18 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139559728","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：