|
on Big Data |
| By: | Abdul Jabbar; Syed Qaisar Jalil |
| Abstract: | This study evaluates the performance of 41 machine learning models, including 21 classifiers and 20 regressors, in predicting Bitcoin prices for algorithmic trading. By examining these models under various market conditions, we highlight their accuracy, robustness, and adaptability to the volatile cryptocurrency market. Our comprehensive analysis reveals the strengths and limitations of each model, providing critical insights for developing effective trading strategies. We employ both machine learning metrics (e.g., Mean Absolute Error, Root Mean Squared Error) and trading metrics (e.g., Profit and Loss percentage, Sharpe Ratio) to assess model performance. Our evaluation includes backtesting on historical data, forward testing on recent unseen data, and real-world trading scenarios, ensuring the robustness and practical applicability of our models. Key findings demonstrate that certain models, such as Random Forest and Stochastic Gradient Descent, outperform others in terms of profit and risk management. These insights offer valuable guidance for traders and researchers aiming to leverage machine learning for cryptocurrency trading. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.18334 |
| By: | Wolff, Dominik; Echterling, Fabian |
| Abstract: | We analyze machine learning algorithms for stock selection. Our study builds on weekly data for the historical constituents of the S&P500 over the period from January 1999 to March 2021 and builds on typical equity factors, additional firm fundamentals, and technical indicators. A variety of machine learning models are trained on the binary classification task to predict whether a specific stock outperforms or underperforms the cross‐sectional median return over the subsequent week. We analyze weekly trading strategies that invest in stocks with the highest predicted outperformance probability. Our empirical results show substantial and significant outperformance of machine learning‐based stock selection models compared to an equally weighted benchmark. Interestingly, we find more simplistic regularized logistic regression models to perform similarly well compared to more complex machine learning models. The results are robust when applied to the STOXX Europe 600 as alternative asset universe. |
| Date: | 2024–01 |
| URL: | https://d.repec.org/n?u=RePEc:dar:wpaper:149079 |
| By: | Undral Byambadalai; Tatsushi Oka; Shota Yasui |
| Abstract: | We propose a novel regression adjustment method designed for estimating distributional treatment effect parameters in randomized experiments. Randomized experiments have been extensively used to estimate treatment effects in various scientific fields. However, to gain deeper insights, it is essential to estimate distributional treatment effects rather than relying solely on average effects. Our approach incorporates pre-treatment covariates into a distributional regression framework, utilizing machine learning techniques to improve the precision of distributional treatment effect estimators. The proposed approach can be readily implemented with off-the-shelf machine learning methods and remains valid as long as the nuisance components are reasonably well estimated. Also, we establish the asymptotic properties of the proposed estimator and present a uniformly valid inference method. Through simulation results and real data analysis, we demonstrate the effectiveness of integrating machine learning techniques in reducing the variance of distributional treatment effect estimators in finite samples. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.16037 |
| By: | Alejandra de la Rica Escudero; Eduardo C. Garrido-Merchan; Maria Coronado-Vaca |
| Abstract: | Financial portfolio management investment policies computed quantitatively by modern portfolio theory techniques like the Markowitz model rely on a set on assumptions that are not supported by data in high volatility markets. Hence, quantitative researchers are looking for alternative models to tackle this problem. Concretely, portfolio management is a problem that has been successfully addressed recently by Deep Reinforcement Learning (DRL) approaches. In particular, DRL algorithms train an agent by estimating the distribution of the expected reward of every action performed by an agent given any financial state in a simulator. However, these methods rely on Deep Neural Networks model to represent such a distribution, that although they are universal approximator models, they cannot explain its behaviour, given by a set of parameters that are not interpretable. Critically, financial investors policies require predictions to be interpretable, so DRL agents are not suited to follow a particular policy or explain their actions. In this work, we developed a novel Explainable Deep Reinforcement Learning (XDRL) approach for portfolio management, integrating the Proximal Policy Optimization (PPO) with the model agnostic explainable techniques of feature importance, SHAP and LIME to enhance transparency in prediction time. By executing our methodology, we can interpret in prediction time the actions of the agent to assess whether they follow the requisites of an investment policy or to assess the risk of following the agent suggestions. To the best of our knowledge, our proposed approach is the first explainable post hoc portfolio management financial policy of a DRL agent. We empirically illustrate our methodology by successfully identifying key features influencing investment decisions, which demonstrate the ability to explain the agent actions in prediction time. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.14486 |
| By: | Höschle, Lisa; Yu, Xiaohua |
| Keywords: | Agribusiness |
| Date: | 2023–09–01 |
| URL: | https://d.repec.org/n?u=RePEc:ags:gewi23:344249 |
| By: | Tiago Monteiro |
| Abstract: | In the field of quantitative finance, machine learning methods have become essential for alpha generation. This paper presents a pioneering method that uniquely combines Hidden Markov Models (HMM) and neural networks, creating a dual-model alpha generation system integrated with Black-Litterman portfolio optimization. The methodology, implemented on the QuantConnect platform, aims to predict future price movements and optimize trading strategies. Specifically, it filters for highly liquid, top-cap energy stocks to ensure stable and predictable performance while also accounting for broker payments. QuantConnect was selected because of its robust framework and to guarantee experimental reproducibility. The algorithm achieved a 31% return between June 1, 2023, and January 1, 2024, with a Sharpe ratio of 1.669, demonstrating its potential. The findings suggest significant improvements in trading strategy performance through the combined use of the HMM and neural networks. This study explores the architecture of the algorithm, data pre-processing techniques, model training procedures, and performance evaluation, highlighting its practical applicability and effectiveness in real-world trading environments. The full code and backtesting data are available under the MIT license. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.19858 |
| By: | Andrew Green |
| Abstract: | Most workers who will be exposed to artificial intelligence (AI) will not require specialised AI skills (e.g. machine learning, natural language processing, etc.). Even so, AI will change the tasks these workers do, and the skills they require. This report provides first estimates for the effect of artificial intelligence on the demand for skills in jobs that do not require specialised AI skills. The results show that the skills most demanded in occupations highly exposed to AI are management and business skills. These include skills in general project management, finance, administration and clerical tasks. The results also show that there have been increases over time in the demand for these skills in occupations highly exposed to AI. For example, the share of vacancies in these occupations that demand at least one emotional, cognitive or digital skill has increased by 8 percentage points. However, using a panel of establishments (which induces plausibly exogenous variation in AI exposure), the report finds evidence that the demand for these skills is beginning to fall. |
| Keywords: | Artificial intelligence, Labour demand, Skills |
| JEL: | J23 J24 J63 |
| Date: | 2024–04–10 |
| URL: | https://d.repec.org/n?u=RePEc:oec:comaaa:14-en |
| By: | Andrew Green |
| Abstract: | Most workers who will be exposed to artificial intelligence (AI) will not require specialised AI skills (e.g. machine learning, natural language processing, etc.). Even so, AI will change the tasks these workers do, and the skills they require. This report provides first estimates for Canada on the effect of artificial intelligence on the demand for skills in jobs that do not require specialised AI skills. The results show that the skills most demanded in occupations highly exposed to AI are management, communication and digital skills. These include skills in budgeting, accounting, written communication, as well as competencies in basic word processing and spreadsheet software. The results also show that, in Canada, demand for social and language skills have increased the most over the past decade in occupations highly exposed to AI. Using a panel of establishments confirms the increasing demand for social and language skills, as well as rising demand for production and physical skills, which may be complementary to AI. However, the establishment panel also finds evidence of decreasing demand for business, management and digital skills in establishments more exposed to AI. |
| Keywords: | Artificial Intelligence, Canada, Skills |
| JEL: | J23 J24 J63 |
| Date: | 2024–05–30 |
| URL: | https://d.repec.org/n?u=RePEc:oec:comaaa:17-en |
| By: | Taha Barwahwala; Aprajit Mahajan; Shekhar Mittal; Ofir Reich |
| Abstract: | We investigate the use of a machine learning (ML) algorithm to identify fraudulent non-existent firms. Using a rich dataset of tax returns over several years in an Indian state, we train an ML-based model to predict fraudulent firms. We then use the model predictions to carry out field inspections of firms identified as suspicious by the ML tool. We find that the ML model is accurate in both simulated and field settings in identifying non-existent firms. Withholding a randomly selected group of firms from inspection, we estimate the causal impact of ML driven inspections. Despite its strong predictive and field performance, the model driven inspections do not yield a significant increase in enforcement as measured by the cancellation of fraudulent firm registrations and tax recovery. We provide two rationales for this discrepancy based on a close analysis of the tax department’s operating protocols: selection bias, and institutional friction in integrating the model into existing administrative systems. Our study serves as a cautionary tale for the application of machine learning in public policy contexts and of relying solely on test set performance as an effectiveness indicator. Field evaluations are critical in assessing the real-world impact of predictive models. |
| JEL: | H0 O10 |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:32705 |
| By: | Shi, Chengchun; Qi, Zhengling; Wang, Jianing; Zhou, Fan |
| Abstract: | Reinforcement learning (RL) is a powerful machine learning technique that enables an intelligent agent to learn an optimal policy that maximizes the cumulative rewards in sequential decision making. Most of methods in the existing literature are developed in online settings where the data are easy to collect or simulate. Motivated by high stake domains such as mobile health studies with limited and pre-collected data, in this article, we study offline reinforcement learning methods. To efficiently use these datasets for policy optimization, we propose a novel value enhancement method to improve the performance of a given initial policy computed by existing state-of-the-art RL algorithms. Specifically, when the initial policy is not consistent, our method will output a policy whose value is no worse and often better than that of the initial policy. When the initial policy is consistent, under some mild conditions, our method will yield a policy whose value converges to the optimal one at a faster rate than the initial policy, achieving the desired“value enhancement” property. The proposed method is generally applicable to any parameterized policy that belongs to certain pre-specified function class (e.g., deep neural networks). Extensive numerical studies are conducted to demonstrate the superior performance of our method. Supplementary materials for this article are available online. |
| Keywords: | mobile health studies; offline reinforcement learning; semi-parametric efficiency; trust region optimization; National Natural Science Founda-tion of China (12001356 |
| JEL: | C1 |
| Date: | 2023–07–20 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:122756 |
| By: | Wüpper, David; Oluoch, Wyclife Agumba; Hadi |
| Abstract: | Agricultural and environmental economists are in the fortunate position that a lot of what is happening on the ground is observable from space. Most agricultural production happens in the open and one can see from space when and where innovations are adopted, crop yields change, or forests are converted to pastures, to name just a few examples. However, converting images into measurements of a particular variable is not trivial, as there are more pitfalls and nuances than “meet the eye”. Overall, however, research benefits tremendously from advances in available satellite data as well as complementary tools, such as cloud-based platforms for data processing, and machine learning algorithms to detect phenomena and mapping variables. The focus of this keynote is to provide agricultural and environmental economists with an accessible introduction to working with satellite data, show-case applications, discuss advantages and weaknesses of satellite data, and emphasize best practices. This is supported by extensive Supplementary Materials, explaining the technical foundations, describing in detail how to create different variables, sketch out work flows, and a discussion of required resources and skills. Last but not least, example data and reproducible codes are available online. |
| Keywords: | Environmental Economics and Policy, Research Methods/ Statistical Methods |
| Date: | 2024–07–26 |
| URL: | https://d.repec.org/n?u=RePEc:ags:cfcp15:344359 |
| By: | Ludovic Goudenege; Andrea Molent; Antonino Zanette |
| Abstract: | This paper explores the application of Machine Learning techniques for pricing high-dimensional options within the framework of the Uncertain Volatility Model (UVM). The UVM is a robust framework that accounts for the inherent unpredictability of market volatility by setting upper and lower bounds on volatility and the correlation among underlying assets. By leveraging historical data and extreme values of estimated volatilities and correlations, the model establishes a confidence interval for future volatility and correlations, thus providing a more realistic approach to option pricing. By integrating advanced Machine Learning algorithms, we aim to enhance the accuracy and efficiency of option pricing under the UVM, especially when the option price depends on a large number of variables, such as in basket or path-dependent options. Our approach evolves backward in time, dynamically selecting at each time step the most expensive volatility and correlation for each market state. Specifically, it identifies the particular values of volatility and correlation that maximize the expected option value at the next time step. This is achieved through the use of Gaussian Process regression, the computation of expectations via a single step of a multidimensional tree and the Sequential Quadratic Programming optimization algorithm. The numerical results demonstrate that the proposed approach can significantly improve the precision of option pricing and risk management strategies compared with methods already in the literature, particularly in high-dimensional contexts. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.13213 |
| By: | Kerri Lu; Stephen Bates; Sherrie Wang |
| Abstract: | Remote sensing map products are used to obtain estimates of environmental quantities, such as deforested area or the effect of conservation zones on deforestation. However, the quality of map products varies, and - because maps are outputs of complex machine learning algorithms that take in a variety of remotely sensed variables as inputs - errors are difficult to characterize. Without capturing the biases that may be present, naive calculations of population-level estimates from such maps are statistically invalid. In this paper, we compare several uncertainty quantification methods - stratification, Olofsson area estimation method, and prediction-powered inference - that combine a small amount of randomly sampled ground truth data with large-scale remote sensing map products to generate statistically valid estimates. Applying these methods across four remote sensing use cases in area and regression coefficient estimation, we find that they result in estimates that are more reliable than naively using the map product as if it were 100% accurate and have lower uncertainty than using only the ground truth and ignoring the map product. Prediction-powered inference uses ground truth data to correct for bias in the map product estimate and (unlike stratification) does not require us to choose a map product before sampling. This is the first work to (1) apply prediction-powered inference to remote sensing estimation tasks, and (2) perform uncertainty quantification on remote sensing regression coefficients without assumptions on the structure of map product errors. To improve the utility of machine learning-generated remote sensing maps for downstream applications, we recommend that map producers provide a holdout ground truth dataset to be used for calibration in uncertainty quantification alongside their maps. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.13659 |
| By: | Francesca Borgonovi; Flavio Calvino; Chiara Criscuolo; Lea Samek; Helke Seitz; Julia Nania; Julia Nitschke; Layla O’Kane |
| Abstract: | This report analyses the demand for positions that require skills needed to develop or work with AI systems across 14 OECD countries between 2019 and 2022. It finds that, despite rapid growth in the demand for AI skills, AI-related online vacancies comprised less than 1% of all job postings and were predominantly found in sectors such as ICT and Professional Services. Skills related to Machine Learning were the most sought after. The US-focused part of the study reveals a consistent demand for socio-emotional, foundational, and technical skills across all AI employers. However, leading firms – those who posted the most AI jobs – exhibited a higher demand for AI professionals combining technical expertise with leadership, innovation, and problem-solving skills, underscoring the importance of these competencies in the AI field. |
| Keywords: | Artificial Intelligence, Online vacancies, Skills |
| JEL: | C81 J23 J24 O33 |
| Date: | 2023–10–17 |
| URL: | https://d.repec.org/n?u=RePEc:oec:comaaa:2-en |
| By: | Wenbo Yan; Ying Tan |
| Abstract: | Recently, the incorporation of both temporal features and the correlation across time series has become an effective approach in time series prediction. Spatio-Temporal Graph Neural Networks (STGNNs) demonstrate good performance on many Temporal-correlation Forecasting Problem. However, when applied to tasks lacking periodicity, such as stock data prediction, the effectiveness and robustness of STGNNs are found to be unsatisfactory. And STGNNs are limited by memory savings so that cannot handle problems with a large number of nodes. In this paper, we propose a novel approach called the Temporal-Correlation Graph Pre-trained Network (TCGPN) to address these limitations. TCGPN utilize Temporal-correlation fusion encoder to get a mixed representation and pre-training method with carefully designed temporal and correlation pre-training tasks. Entire structure is independent of the number and order of nodes, so better results can be obtained through various data enhancements. And memory consumption during training can be significantly reduced through multiple sampling. Experiments are conducted on real stock market data sets CSI300 and CSI500 that exhibit minimal periodicity. We fine-tune a simple MLP in downstream tasks and achieve state-of-the-art results, validating the capability to capture more robust temporal correlation patterns. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.18519 |
| By: | Kamesh Korangi; Christophe Mues; Cristi\'an Bravo |
| Abstract: | Apart from assessing individual asset performance, investors in financial markets also need to consider how a set of firms performs collectively as a portfolio. Whereas traditional Markowitz-based mean-variance portfolios are widespread, network-based optimisation techniques have built upon these developments. However, most studies do not contain firms at risk of default and remove any firms that drop off indices over a certain time. This is the first study to incorporate risky firms and use all the firms in portfolio optimisation. We propose and empirically test a novel method that leverages Graph Attention networks (GATs), a subclass of Graph Neural Networks (GNNs). GNNs, as deep learning-based models, can exploit network data to uncover nonlinear relationships. Their ability to handle high-dimensional features and accommodate customised layers for specific purposes makes them particularly appealing for large-scale problems such as mid- and small-cap portfolio optimization. This study utilises 30 years of data on mid-cap firms, creating graphs of firms using distance correlation and the Triangulated Maximally Filtered Graph approach. These graphs are the inputs to a GAT model that we train using custom layers which impose weight and allocation constraints and a loss function derived from the Sharpe ratio, thus directly maximising portfolio risk-adjusted returns. This new model is benchmarked against a network characteristic-based portfolio, a mean variance-based portfolio, and an equal-weighted portfolio. The results show that the portfolio produced by the GAT-based model outperforms all benchmarks and is consistently superior to other strategies over a long period while also being informative of market dynamics. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.15532 |
| By: | Tian Guo; Emmanuel Hauptmann |
| Abstract: | Large language models (LLMs) and their fine-tuning techniques have demonstrated superior performance in various language understanding and generation tasks. This paper explores fine-tuning LLMs for stock return forecasting with financial newsflow. In quantitative investing, return forecasting is fundamental for subsequent tasks like stock picking, portfolio optimization, etc. We formulate the model to include text representation and forecasting modules. We propose to compare the encoder-only and decoder-only LLMs, considering they generate text representations in distinct ways. The impact of these different representations on forecasting performance remains an open question. Meanwhile, we compare two simple methods of integrating LLMs' token-level representations into the forecasting module. The experiments on real news and investment universes reveal that: (1) aggregated representations from LLMs' token-level embeddings generally produce return predictions that enhance the performance of long-only and long-short portfolios; (2) in the relatively large investment universe, the decoder LLMs-based prediction model leads to stronger portfolios, whereas in the small universes, there are no consistent winners. Among the three LLMs studied (DeBERTa, Mistral, Llama), Mistral performs more robustly across different universes; (3) return predictions derived from LLMs' text representations are a strong signal for portfolio construction, outperforming conventional sentiment scores. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.18103 |
| By: | St\'ephane Cr\'epey (UFR Math\'ematiques UPCit\'e); Botao Li (LPSM); Hoang Nguyen (IES, LPSM); Bouazza Saadeddine |
| Abstract: | We present a unified framework for computing CVA sensitivities, hedging the CVA, and assessing CVA risk, using probabilistic machine learning meant as refined regression tools on simulated data, validatable by low-cost companion Monte Carlo procedures. Various notions of sensitivities are introduced and benchmarked numerically. We identify the sensitivities representing the best practical trade-offs in downstream tasks including CVA hedging and risk assessment. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.18583 |
| By: | Joel P. Villarino; \'Alvaro Leitao |
| Abstract: | The present work addresses the challenge of training neural networks for Dynamic Initial Margin (DIM) computation in counterparty credit risk, a task traditionally burdened by the high costs associated with generating training datasets through nested Monte Carlo (MC) simulations. By condensing the initial market state variables into an input vector, determined through an interest rate model and a parsimonious parameterization of the current interest rate term structure, we construct a training dataset where labels are noisy but unbiased DIM samples derived from single MC paths. A multi-output neural network structure is employed to handle DIM as a time-dependent function, facilitating training across a mesh of monitoring times. The methodology offers significant advantages: it reduces the dataset generation cost to a single MC execution and parameterizes the neural network by initial market state variables, obviating the need for repeated training. Experimental results demonstrate the approach's convergence properties and robustness across different interest rate models (Vasicek and Hull-White) and portfolio complexities, validating its general applicability and efficiency in more realistic scenarios. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.16435 |
| By: | Arnone, Massimo; Leogrande, Angelo |
| Abstract: | The competitiveness of financed intermediaries cannot be based exclusively on financial sustainability, i.e. the ability to create profit, but it is also necessary to acquire a transversal vision of sustainability focused on the three ESG dimensions. The paper intends to propose a reflection on the main impacts of the integration of ESG factors on business decisionmaking and operational processes in the financial sector. In this context, we try to understand what role FinTech can play in favor of greater sustainability. Furthermore, through an empirical analysis, some determinants relating to social, environmental, and governance issues are identified which influence the volume of financial resources moved in the factoring market at a European level. Machine learning models are also proposed to estimate the volume |
| Keywords: | Sustainability, Factoring, ESG, FinTech, Machine Learning, Clusterization |
| JEL: | G00 G21 G22 |
| Date: | 2024–06–28 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:121342 |
| By: | Bonelli, Maxime (London Business School - Department of Finance); Foucault, Thierry (HEC Paris) |
| Abstract: | Big data allows active asset managers to find new trading signals but doing so requires new skills. Thus, it can reduce the ability of asset managers lacking these skills to produce superior returns. Consistent with this hypothesis, we find that the release of satellite imagery data tracking firms’ parking lots reduces active mutual funds’ stock picking abilities in stocks covered by this data. This decline is stronger for funds that are more likely to rely on traditional sources of expertise (e.g., specialized industry knowledge) to generate their signals, leading them to divest from covered stocks. These results suggest that big data has the potential to displace high-skill workers in finance. |
| Keywords: | Big data; active mutual funds; stock-picking skill; quantitative investment |
| JEL: | G11 G14 G23 |
| Date: | 2023–08–02 |
| URL: | https://d.repec.org/n?u=RePEc:ebg:heccah:1491 |
| By: | Ma, Tao; Yang, Xuzhi; Szabo, Zoltan |
| Abstract: | Reinforcement learning (RL) -- finding the optimal behaviour (also referred to as policy) maximizing the collected long-term cumulative reward -- is among the most influential approaches in machine learning with a large number of successful applications. In several decision problems, however, one faces the possibility of policy switching -- changing from the current policy to a new one -- which incurs a non-negligible cost (examples include the shifting of the currently applied educational technology, modernization of a computing cluster, and the introduction of a new webpage design), and in the decision one is limited to using historical data without the availability for further online interaction. Despite the inevitable importance of this offline learning scenario, to our best knowledge, very little effort has been made to tackle the key problem of balancing between the gain and the cost of switching in a flexible and principled way. Leveraging ideas from the area of optimal transport, we initialize the systematic study of policy switching in offline RL. We establish fundamental properties and design a Net Actor-Critic algorithm for the proposed novel switching formulation. Numerical experiments demonstrate the efficiency of our approach on multiple benchmarks of the Gymnasium. |
| JEL: | C1 |
| Date: | 2024–07–01 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:124144 |
| By: | Hongshen Yang; Avinash Malik |
| Abstract: | Cryptocurrency is a cryptography-based digital asset with extremely volatile prices. Around $70 billion worth of crypto-currency is traded daily on exchanges. Trading crypto-currency is difficult due to the inherent volatility of the crypto-market. In this work, we want to test the hypothesis: "Can techniques from artificial intelligence help with algorithmically trading cryptocurrencies?". In order to address this question, we combine Reinforcement Learning (RL) with pair trading. Pair trading is a statistical arbitrage trading technique which exploits the price difference between statistically correlated assets. We train reinforcement learners to determine when and how to trade pairs of cryptocurrencies. We develop new reward shaping and observation/action spaces for reinforcement learning. We performed experiments with the developed reinforcement learner on pairs of BTC-GBP and BTC-EUR data separated by 1-minute intervals (n = 263, 520). The traditional non-RL pair trading technique achieved an annualised profit of 8.33%, while the proposed RL-based pair trading technique achieved annualised profits from 9.94% - 31.53%, depending upon the RL learner. Our results show that RL can significantly outperform manual and traditional pair trading techniques when applied to volatile markets such as cryptocurrencies. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.16103 |
| By: | Chung I Lu; Julian Sester |
| Abstract: | Generating synthetic financial time series data that accurately reflects real-world market dynamics holds tremendous potential for various applications, including portfolio optimization, risk management, and large scale machine learning. We present an approach for training generative models for financial time series using the maximum mean discrepancy (MMD) with a signature kernel. Our method leverages the expressive power of the signature transform to capture the complex dependencies and temporal structures inherent in financial data. We employ a moving average model to model the variance of the noise input, enhancing the model's ability to reproduce stylized facts such as volatility clustering. Through empirical experiments on S&P 500 index data, we demonstrate that our model effectively captures key characteristics of financial time series and outperforms a comparable GAN-based approach. In addition, we explore the application of the synthetic data generated to train a reinforcement learning agent for portfolio management, achieving promising results. Finally, we propose a method to add robustness to the generative model by tweaking the noise input so that the generated sequences can be adjusted to different market environments with minimal data. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.19848 |
| By: | Beckert, Jens; Arndt, H. Lukas R. |
| Abstract: | Between 2009 and 2015 Greece underwent a profound sovereign debt crisis that led to a serious political crisis in Europe and the restructuring of Greek debt. We argue that the prevalence of negative narratives about the future contributed to the changes in spreads of Greek bonds during the crisis. We support our argument by presenting results from text mining a corpus of 9, 435 articles from the Financial Times and the Wall Street Journal. Based on sentiments and a machine learning model predicting future reference, we identify newspaper articles which generate negative and uncertain outlooks for the future in the expert discourse. We provide evidence from time series regression analysis showing that these negative imagined futures have explanatory power in models estimating spread development of Greek vs. German sovereign bonds. We suggest that these findings provide good evidence for the relevance of "imagined futures" for investors' behavior, and give directions for an innovative contribution of sociology to understanding the microfoundations of financial crises. |
| Abstract: | Zwischen 2009 und 2015 durchlebte Griechenland eine tiefgreifende Staatsschuldenkrise, die zu einer schweren politischen Krise in Europa und zur Umstrukturierung der griechischen Schulden führte. Wir argumentieren, dass die Prävalenz negativer Narrative über die Zukunft zu den Veränderungen der Spreads griechischer Anleihen während der Krise beigetragen hat. Zur Untermauerung dieser These präsentieren wir die Ergebnisse der Textanalyse eines Korpus von 9.435 Artikeln aus der Financial Times und dem Wall Street Journal. Auf der Grundlage von Sentiments und einem maschinellen Lernmodell zur Erkennung von Zukunftsvorhersagen identifizieren wir Zeitungsartikel, die negative und unsichere Zukunftsaussichten im Expertendiskurs erzeugen. Wir zeigen anhand von Zeitreihen-Regressionsanalysen, dass diese negativen Zukunftsvorstellungen Erklärungskraft in Modellen zur Schätzung der Spread-Entwicklung von griechischen gegenüber deutschen Staatsanleihen haben. Diese Ergebnisse liefern Evidenz für die Relevanz imaginierter Zukünfte für das Verhalten von Anlegern und ermöglichen einen innovativen Beitrag der Soziologie zum Verständnis der Mikroebene von Finanzkrisen. |
| Keywords: | bond spreads, economic sociology, financial markets, Greek debt crisis, imagined futures, sentiment analysis, sovereign debt, valuation, Anleihen-Spreads, Bewertung, Finanzmärkte, griechische Schuldenkrise, imaginierte Zukünfte, Staatsverschuldung, Sentimentanalyse, Wirtschaftssoziologie |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:mpifgd:300665 |
| By: | Li, Jie; Fearnhead, Paul; Fryzlewicz, Piotr; Wang, Tengyao |
| Abstract: | Detecting change points in data is challenging because of the range of possible types of change and types of behaviour of data when there is no change. Statistically efficient methods for detecting a change will depend on both of these features, and it can be difficult for a practitioner to develop an appropriate detection method for their application of interest. We show how to automatically generate new offline detection methods based on training a neural network. Our approach is motivated by many existing tests for the presence of a change point being representable by a simple neural network, and thus a neural network trained with sufficient data should have performance at least as good as these methods. We present theory that quantifies the error rate for such an approach, and how it depends on the amount of training data. Empirical results show that, even with limited training data, its performance is competitive with the standard cumulative sum (CUSUM) based classifier for detecting a change in mean when the noise is independent and Gaussian, and can substantially outperform it in the presence of auto-correlated or heavy-tailed noise. Our method also shows strong results in detecting and localizing changes in activity based on accelerometer data. |
| Keywords: | automatic statistician; classification; likelihood-free inference; neural networks; structural breaks; supervised learning; e High End Computing Cluster at Lancaster University; and EPSRC grants EP/V053590/1; EP/V053639/1 and EP/T02772X/1 |
| JEL: | C1 |
| Date: | 2024–04–01 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:120083 |
| By: | Mahdi Ebrahimi Kahou (Bowdoin College); Jesus Fernandez-Villaverde (University of Pennsylvania, NBER, and CEPR); Sebastian Gomez-Cardona (Morningstar); Jesse Perla (University of British Columbia); Jan Rosa (University of British Columbia) |
| Abstract: | In the long run, we are all dead. Nonetheless, when studying the short-run dynamics of economic models, it is crucial to consider boundary conditions that govern long-run, forwardlooking behavior, such as transversality conditions. We demonstrate that machine learning (ML) can automatically satisfy these conditions due to its inherent inductive bias toward finding flat solutions to functional equations. This characteristic enables ML algorithms to solve for transition dynamics, ensuring that long-run boundary conditions are approximately met. ML can even select the correct equilibria in cases of steady-state multiplicity. Additionally, the inductive bias provides a foundation for modeling forward-looking behavioral agents with self-consistent expectations. |
| Keywords: | Machine learning, inductive bias, rational expectations, transitional dynamics, transversality, behavioral macroeconomics |
| JEL: | C1 E1 |
| Date: | 2024–08–12 |
| URL: | https://d.repec.org/n?u=RePEc:pen:papers:24-019 |
| By: | Foltas, Alexander |
| Abstract: | I contribute to previous research on the efficient integration of forecasters' narratives into business cycle forecasts. Using a Bidirectional Encoder Representations from Transformers (BERT) model, I quantify 19, 300 paragraphs from German business cycle reports (1998-2021) and classify the signs of institutes' consumption forecast errors. The correlation is strong for 12.8% of paragraphs with a predicted class probability of 85% or higher. Reviewing 150 of such high-probability paragraphs reveals recurring narratives. Underestimations of consumption growth often mention rising employment, increasing wages and transfer payments, low inflation, decreasing taxes, crisis-related fiscal support, and reduced relevance of marginal employment. Conversely, overestimated consumption forecasts present opposing narratives. Forecasters appear to particularly underestimate these factors when they disproportionately affect low-income households. |
| Keywords: | Macroeconomic forecasting, Evaluating forecasts, Business cycles, Consumption forecasting, Natural language processing, Language Modeling, Machine learning, Judgemental forecasting |
| JEL: | E21 C53 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:pp1859:300847 |
| By: | Abe C. Dunn; Eric English; Kyle K. Hood; Lowell Mason; Brian Quistorff |
| Abstract: | Big data offers potentially enormous benefits for improving economic measurement, but it also presents challenges (e.g., lack of representativeness and instability), implying that their value is not always clear. We propose a framework for quantifying the usefulness of these data sources for specific applications, relative to existing official sources. We specifically weigh the potential benefits of additional granularity and timeliness, while examining the accuracy associated with any new or improved estimates, relative to comparable accuracy produced in existing official statistics. We apply the methodology to employment estimates using data from a payroll processor, considering both the improvement of existing state-level estimates, but also the production of new, more timely, county-level estimates. We find that incorporating payroll data can improve existing state-level estimates by 11\% based on out-of-sample mean absolute error, although the improvement is considerably higher for smaller state-industry cells. We also produce new county-level estimates that could provide more timely granular estimates than previously available. We develop a novel test to determine if these new county-level estimates have errors consistent with official series. Given the level of granularity, we cannot reject the hypothesis that the new county estimates have an accuracy in line with official measures, implying an expansion of the existing frontier. We demonstrate the practical importance of these experimental estimates by investigating a hypothetical application during the COVID-19 pandemic, a period in which more timely and granular information could have assisted in implementing effective policies. Relative to existing estimates, we find that the alternative payroll data series could help identify areas of the country where employment was lagging. Moreover, we also demonstrate the value of a more timely series. |
| JEL: | E01 E24 R11 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:bea:papers:0128 |
| By: | Shi, Chengchun; Zhou, Yunzhe; Li, Lexin |
| Abstract: | In this article, we propose a new hypothesis testing method for directed acyclic graph (DAG). While there is a rich class of DAG estimation methods, there is a relative paucity of DAG inference solutions. Moreover, the existing methods often impose some specific model structures such as linear models or additive models, and assume independent data observations. Our proposed test instead allows the associations among the random variables to be nonlinear and the data to be time-dependent. We build the test based on some highly flexible neural networks learners. We establish the asymptotic guarantees of the test, while allowing either the number of subjects or the number of time points for each subject to diverge to infinity. We demonstrate the efficacy of the test through simulations and a brain connectivity network analysis. Supplementary materials for this article are available online. |
| Keywords: | brain connectivity networks; directed acrylic graph; hypothesis testing; generative adversarial networks; multilayer perceptron neural networks; Hypothesis testing; CIF-2102227; R01AG061303; R01AG062542; EP/W014971/1 |
| JEL: | C1 |
| Date: | 2023–07–12 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:119446 |
| By: | Davillas, Apostolos (University of Macedonia); Jones, Andrew M. (University of York) |
| Abstract: | We explore the role of epigenetic biological age in predicting subsequent health care utilisation. We use longitudinal data from the UK Understanding Society panel, capitalizing on the availability of baseline epigenetic biological age measures along with data on general practitioner (GP) consultations, outpatient (OP) visits, and hospital inpatient (IP) care collected 5-12 years from baseline. Using least absolute shrinkage and selection operator (LASSO) regression analyses and accounting for participants' pre-existing health conditions, baseline biological underlying health, and socio-economic predictors we find that biological age predicts future GP consultations and IP care, while chronological rather than biological age matters for future OP visits. Post-selection prediction analysis and Shapley-Shorrocks decompositions, comparing our preferred prediction models to models that replace biological age with chronological age, suggest that biological ageing has a stronger role in the models predicting future IP care as opposed to "gatekeeping" GP consultations. |
| Keywords: | epigenetics, biological age, health care utilisation, red herring hypothesis, LASSO, supervised machine learning |
| JEL: | C5 C81 I10 I18 |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp17159 |
| By: | Adria Pop; Jan Sp\"orer; Siegfried Handschuh |
| Abstract: | This research dissects financial equity research reports (ERRs) by mapping their content into categories. There is insufficient empirical analysis of the questions answered in ERRs. In particular, it is not understood how frequently certain information appears, what information is considered essential, and what information requires human judgment to distill into an ERR. The study analyzes 72 ERRs sentence-by-sentence, classifying their 4940 sentences into 169 unique question archetypes. We did not predefine the questions but derived them solely from the statements in the ERRs. This approach provides an unbiased view of the content of the observed ERRs. Subsequently, we used public corporate reports to classify the questions' potential for automation. Answers were labeled "text-extractable" if the answers to the question were accessible in corporate reports. 78.7% of the questions in ERRs can be automated. Those automatable question consist of 48.2% text-extractable (suited to processing by large language models, LLMs) and 30.5% database-extractable questions. Only 21.3% of questions require human judgment to answer. We empirically validate using Llama-3-70B and GPT-4-turbo-2024-04-09 that recent advances in language generation and information extraction enable the automation of approximately 80% of the statements in ERRs. Surprisingly, the models complement each other's strengths and weaknesses well. The research confirms that the current writing process of ERRs can likely benefit from additional automation, improving quality and efficiency. The research thus allows us to quantify the potential impacts of introducing large language models in the ERR writing process. The full question list, including the archetypes and their frequency, will be made available online after peer review. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.18327 |
| By: | Shengkun Wang; Taoran Ji; Jianfeng He; Mariam Almutairi; Dan Wang; Linhan Wang; Min Zhang; Chang-Tien Lu |
| Abstract: | Stock volatility prediction is an important task in the financial industry. Recent advancements in multimodal methodologies, which integrate both textual and auditory data, have demonstrated significant improvements in this domain, such as earnings calls (Earnings calls are public available and often involve the management team of a public company and interested parties to discuss the company's earnings). However, these multimodal methods have faced two drawbacks. First, they often fail to yield reliable models and overfit the data due to their absorption of stochastic information from the stock market. Moreover, using multimodal models to predict stock volatility suffers from gender bias and lacks an efficient way to eliminate such bias. To address these aforementioned problems, we use adversarial training to generate perturbations that simulate the inherent stochasticity and bias, by creating areas resistant to random information around the input space to improve model robustness and fairness. Our comprehensive experiments on two real-world financial audio datasets reveal that this method exceeds the performance of current state-of-the-art solution. This confirms the value of adversarial training in reducing stochasticity and bias for stock volatility prediction tasks. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.18324 |
| By: | Gulfam Haider; Laiba Zubair; Aman Saleem |
| Abstract: | This study, rooted in dynamic capability theory and the developing era of Big Data Analytics, explores the transformative effect of BDA EDCs on marketing. Ambidexterity and firms market performance in the textile sector of Pakistans cities. Specifically, focusing on the firms who directly deal with customers, investigates the nuanced role of BDA EDCs in textile retail firms potential to navigate market dynamics. Emphasizing the exploitation component of marketing ambidexterity, the study investigated the mediating function of marketing ambidexterity and the moderating influence of competitive pressure. Using a survey questionnaire, the study targets key choice makers in textile firms of Faisalabad, Chiniot and Lahore, Pakistan. The PLS-SEM model was employed as an analytical technique, allows for a full examination of the complicated relations between BDA EDCs, marketing ambidexterity, rival pressure, and market performance. The study Predicting a positive impact of Big Data on marketing ambidexterity, with a specific emphasis on exploitation. The study expects this exploitation-orientated marketing ambidexterity to significantly enhance the firms market performance. This research contributes to the existing literature on dynamic capabilities-based frameworks from the perspective of the retail segment of textile industry. The study emphasizes the role of BDA-EDCs in the retail sector, imparting insights into the direct and indirect results of BDA EDCs on market performance inside the retail area. The study s novelty lies in its contextualization of BDA-EDCs in the textile zone of Faisalabad, Lahore and Chiniot, providing a unique perspective on the effect of BDA on marketing ambidexterity and market performance in firms. Methodologically, the study uses numerous samples of retail sectors to make sure broader universality, contributing realistic insights. |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2407.15522 |
| By: | Guan-Yuan Wang (Vilnius University, Faculty of Mathematics and Informatics) |
| Abstract: | Many game development companies use game data analysis for mining insights about users' behaviour and possible product growth. One of the most important analysis tasks for game development is user churn prediction. Effective churn prediction can help hold users in the game by initiating additional actions for their engagement. We focused on high-value user churn prediction as it is of particular interest for any business to keep paying customers satisfied and engaged. We consider the churn prediction problem as a classification problem and conduct the random undersampling approach to address imbalanced class distribution between churners and active users. Based on our real-life data from a freemium casual mobile game, although the best model was chosen as the final classification algorithm for extracted data, we can definitely say there is no general solution to the stated problem. Model performance highly depends on the churn definition, user segmentation and feature engineering, it is therefore necessary to have a custom approach to churn analysis in each specific case. |
| Keywords: | Churn prediction, mobile games, classification models, resamlpling methods, imbalanced class distribution, machine learning |
| Date: | 2022–12–16 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-04632443 |
| By: | Yacine Aït-Sahalia; Chen Xu Li; Chenxu Li |
| Abstract: | This paper relates jumps in high frequency stock prices to firm-level, industry and macroeconomic news, in the form of machine-readable releases from Thomson Reuters News Analytics. We find that most relevant news, both idiosyncratic and systematic, lead quickly to price jumps, as market efficiency suggests they should. However, in the reverse direction, the vast majority of price jumps do not have identifiable public news that can explain them, in a departure from the ideal of a fair, orderly and efficient market. Microstructure-driven variables have only limited predictive power to help distinguish between jumps with and without news. |
| JEL: | G12 G14 |
| Date: | 2024–07 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:32746 |