|
on Computational Economics |
| By: | Manuela Pedio; Massimo Guidolin; Giulia Panzeri |
| Abstract: | Machine learning is significantly shaping the advancement of various fields, and among them, notably, finance, where its range of applications and efficiency impacts are seemingly boundless. Contemporary techniques, particularly in reinforcement learning, have prompted both practitioners and academics to contemplate the potential of an artificial intelligence revolution in portfolio management. In this paper, we provide an overview of the primary methods in machine learning currently utilized in portfolio decision-making. We delve into discussions surrounding the existing limitations of machine learning algorithms and explore prevailing hypotheses regarding their future expansions. Specifically, we categorize and analyze the applications of machine learning in systematic trading strategies, portfolio weight optimization, smart beta and passive investment strategies, textual analysis, and trade execution, each separately surveyed for a comprehensive understanding. |
| Keywords: | Machine learning; portfolio choice; artificial intelligence; neural language processing; stock return predictions, market timing, mean-variance asset allocation. |
| JEL: | C45 C61 G10 G11 G17 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:baf:cbafwp:cbafwp24233 |
| By: | Diego Vallarino |
| Abstract: | This paper introduces a novel approach to optimizing portfolio rebalancing by integrating Graph Neural Networks (GNNs) for predicting transaction costs and Dijkstra's algorithm for identifying cost-efficient rebalancing paths. Using historical stock data from prominent technology firms, the GNN is trained to forecast future transaction costs, which are then applied as edge weights in a financial asset graph. Dijkstra's algorithm is used to find the least costly path for reallocating capital between assets. Empirical results show that this hybrid approach significantly reduces transaction costs, offering a powerful tool for portfolio managers, especially in high-frequency trading environments. This methodology demonstrates the potential of combining advanced machine learning techniques with classical optimization algorithms to improve financial decision-making processes. Future research will explore expanding the asset universe and incorporating reinforcement learning for continuous portfolio optimization. |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2410.01864 |
| By: | Leite, Walter; Zhang, Huibin; collier, zachary; Chawla, Kamal; , l.kong@ufl.edu; Lee, Yongseok (University of Florida); Quan, Jia; Soyoye, Olushola |
| Abstract: | Machine learning has become a common approach for estimating propensity scores for quasi-experimental research using matching, weighting, or stratification on the propensity score. This systematic review examined machine learning applications for propensity score estimation across different fields, such as health, education, social sciences, and business over 40 years. The results show that the gradient boosting machine (GBM) is the most frequently used method, followed by random forest. Classification and regression trees (CART), neural networks, and the super learner were also used in more than five percent of studies. The most frequently used packages to estimate propensity scores were twang, gbm and randomforest in the R statistical software. The review identified many hyperparameter configurations used for machine learning methods. However, it also shows that hyperparameters are frequently under-reported, as well as critical steps of the propensity score analysis, such as the covariate balance evaluation. A set of guidelines for reporting the use of machine learning for propensity score estimation is provided. |
| Date: | 2024–10–09 |
| URL: | https://d.repec.org/n?u=RePEc:osf:osfxxx:gmrk7 |
| By: | Dupret, Jean-Loup (Université catholique de Louvain, LIDAM/ISBA, Belgium); Hainaut, Donatien (Université catholique de Louvain, LIDAM/ISBA, Belgium) |
| Abstract: | This paper introduces the GPI-PINN algorithm, a novel numerical scheme for solving continuous-time stochastic optimal control problems in high dimensions when the optimal control does not admit an explicit solution. Combining Physics-Informed Neural Networks with an Actor-Critic structure built upon the Generalized Policy Iteration technique, this successive deep learning algorithm employs two separate neural networks to approximate both the value function and the multidimensional optimal control. This way, the GPI-PINN algorithm manages to achieve a global approximation of the optimal solution across all time and space, which can be evaluated online rapidly. The optimality and convergence of the scheme are demonstrated theoretically and its accuracy and efficacy are shown empirically based on two numerical examples. In particular, we generalize the standard Almgren-Chriss model arising from optimal liquidation in finance by allowing for a price impact model with fully nonlinear temporary and permanent impact functions and by considering a multidimensional setting with numerous co-integrated assets. |
| Keywords: | Machine learning ; Stochastic optimal control ; Deep learning ; Physics-Informed Neural Networks ; Optimal liquidation |
| Date: | 2024–05–30 |
| URL: | https://d.repec.org/n?u=RePEc:aiz:louvad:2024016 |
| By: | Zongwu Cai (Department of Economics, The University of Kansas, Lawrence, KS 66045, USA); Pixiong Chen (Division of Model Risk Management, Wells Fargo Bank, Charlotte, NC 28202, USA) |
| Abstract: | In this paper, we propose utilizing machine learning methods to determine the expected aggregated stock market risk premium based on online investor sentiment and employing the multifold forward-validation method to select the relevant hyperparameters. Our empirical studies provide strong evidence that some machine learning methods, such as extreme gradient boosting or random forest, show significant predictive ability in terms of their out-of-sample performances with high-dimensional investor sentiment proxies. They also outperform the traditional linear models, which shows a possible unobserved nonlinear relationship between online investor sentiment and risk premium. Moreover, this predictability based on online investor sentiment has a better economic value, so it improves portfolio performance for investors who need to decide the optimal asset allocation in terms of the certainty equivalent return gain and the Sharpe ratio. |
| Keywords: | Asset return; Machine learning; Nonlinearity; Portfolio allocations; Predictability. |
| JEL: | C45 C55 C58 G11 G17 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:kan:wpaper:202411 |
| By: | Yanxin Shen; Pulin Kirin Zhang |
| Abstract: | Financial sentiment analysis (FSA) is crucial for evaluating market sentiment and making well-informed financial decisions. The advent of large language models (LLMs) such as BERT and its financial variant, FinBERT, has notably enhanced sentiment analysis capabilities. This paper investigates the application of LLMs and FinBERT for FSA, comparing their performance on news articles, financial reports and company announcements. The study emphasizes the advantages of prompt engineering with zero-shot and few-shot strategy to improve sentiment classification accuracy. Experimental results indicate that GPT-4o, with few-shot examples of financial texts, can be as competent as a well fine-tuned FinBERT in this specialized field. |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2410.01987 |
| By: | Liu, Yirui; Qiao, Xinghao; Pei, Yulong; Wang, Liying |
| Abstract: | This paper introduces the Deep Functional Factor Model (DF2M), a Bayesian nonparametric model designed for analysis of high-dimensional functional time series. DF2M is built upon the Indian Buffet Process and the multi-task Gaussian Process, incorporating a deep kernel function that captures non-Markovian and nonlinear temporal dynamics. Unlike many black-box deep learning models, DF2M offers an explainable approach to utilizing neural networks by constructing a factor model and integrating deep neural networks within the kernel function. Additionally, we develop a computationally efficient variational inference algorithm to infer DF2M. Empirical results from four real-world datasets demonstrate that DF2M provides better explainability and superior predictive accuracy compared to conventional deep learning models for high-dimensional functional time series. |
| JEL: | C1 |
| Date: | 2024–07–21 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:125587 |
| By: | Lucas Gomes; Jannis Kueck; Mara Mattes; Martin Spindler; Alexey Zaytsev |
| Abstract: | Collusion is a complex phenomenon in which companies secretly collaborate to engage in fraudulent practices. This paper presents an innovative methodology for detecting and predicting collusion patterns in different national markets using neural networks (NNs) and graph neural networks (GNNs). GNNs are particularly well suited to this task because they can exploit the inherent network structures present in collusion and many other economic problems. Our approach consists of two phases: In Phase I, we develop and train models on individual market datasets from Japan, the United States, two regions in Switzerland, Italy, and Brazil, focusing on predicting collusion in single markets. In Phase II, we extend the models' applicability through zero-shot learning, employing a transfer learning approach that can detect collusion in markets in which training data is unavailable. This phase also incorporates out-of-distribution (OOD) generalization to evaluate the models' performance on unseen datasets from other countries and regions. In our empirical study, we show that GNNs outperform NNs in detecting complex collusive patterns. This research contributes to the ongoing discourse on preventing collusion and optimizing detection methodologies, providing valuable guidance on the use of NNs and GNNs in economic applications to enhance market fairness and economic welfare. |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2410.07091 |
| By: | Holtdirk, Tobias; Assenmacher, Dennis; Bleier, Arnim; Wagner, Claudia |
| Abstract: | Surveys are a cornerstone of empirical social science research, providing invaluable insights into the opinions, beliefs, behaviours, and characteristics of people. However, issues such as refusal to participate, skipping questions, sampling bias, and attrition significantly impact the quality and reliability of survey data. Recently, researchers have started investigating the potential of Large Language Models (LLMs) to role-play a pre-defined set of "characters" and simulate their survey responses with little or no additional training data and costs. While previous research on forecasting, imputing, and simulating survey answers with LLMs has focused on zero-shot and few-shot approaches, this study investigates the viability of fine-tuning LLMs to simulate responses of survey participants. We fine-tune Large Language Models (LLMs) on subsets of the data from the German Longitudinal Election Study (GLES) and evaluate their predictive performance on the "vote choice" for a random set of held-out participants. We compare the LLMs' performance against various baseline methods. Our findings show that small, fine-tuned open-source LLMs can outperform zero-shot predictions of larger LLMs. They are able to match the performance of established tabular data classifiers, are more sample efficient, and outperform them in cases with systematic non-responses. This study contributes to the growing body of research on LLMs for simulating survey data by demonstrating the effectiveness of fine-tuning approaches. |
| Date: | 2024–10–07 |
| URL: | https://d.repec.org/n?u=RePEc:osf:osfxxx:udz28 |
| By: | Yang Liu; Ran Pan; Rui Xu |
| Abstract: | Forecasting inflation has become a major challenge for central banks since 2020, due to supply chain disruptions and economic uncertainty post-pandemic. Machine learning models can improve forecasting performance by incorporating a wider range of variables, allowing for non-linear relationships, and focusing on out-of-sample performance. In this paper, we apply machine learning (ML) models to forecast near-term core inflation in Japan post-pandemic. Japan is a challenging case, because inflation had been muted until 2022 and has now risen to a level not seen in four decades. Four machine learning models are applied to a large set of predictors alongside two benchmark models. For 2023, the two penalized regression models systematically outperform the benchmark models, with LASSO providing the most accurate forecast. Useful predictors of inflation post-2022 include household inflation expectations, inbound tourism, exchange rates, and the output gap. |
| Keywords: | Core inflation; forecasting; machine learning models; LASSO; Japan |
| Date: | 2024–09–27 |
| URL: | https://d.repec.org/n?u=RePEc:imf:imfwpa:2024/206 |
| By: | Diego Vallarino |
| Abstract: | This study evaluates the effectiveness of a Mixture of Experts (MoE) model for stock price prediction by comparing it to a Recurrent Neural Network (RNN) and a linear regression model. The MoE framework combines an RNN for volatile stocks and a linear model for stable stocks, dynamically adjusting the weight of each model through a gating network. Results indicate that the MoE approach significantly improves predictive accuracy across different volatility profiles. The RNN effectively captures non-linear patterns for volatile companies but tends to overfit stable data, whereas the linear model performs well for predictable trends. The MoE model's adaptability allows it to outperform each individual model, reducing errors such as Mean Squared Error (MSE) and Mean Absolute Error (MAE). Future work should focus on enhancing the gating mechanism and validating the model with real-world datasets to optimize its practical applicability. |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2410.07234 |
| By: | Benedikt Gloria; Sven Bienert; Johannes Melsbach; Detlef Schoder |
| Abstract: | In recent times, large language models (LLMs) such as ChatGPT and LLaMA have gained significant attention. These models demonstrate remarkable capability in solving complex tasks, drawing knowledge primarily from a generalized database rather than niche subject areas. Consequently, there has been a growing demand for domain-specific LLMs tailored to social and natural sciences, such as BioGPT or BloombergGPT. In this study, we present our own domain-specific LLM focused on real estate, based on the parameter-efficient finetuning technique known as Low-rank adaptation (LoRA) applied to the Mistral 7B model. To create a comprehensive finetuning dataset, we compiled a curated 21k self-instruction dataset sourced from 670 scientific papers, market research, scholarly articles and real estate books. To assess the efficacy of Real-GPT, we devised a set of ca. 5, 000 multiple-choice questions to gauge the real estate knowledge of the models. Despite its notably compact size, our model outperforms other cutting-edge models. Consequently, our developed model not only showcases superior performance but also illustrates its capacity to facilitate investment decisions, interpret current market data, and potentially simplify property valuation processes. This development showcases the potential of LLMs to revolutionize the field of real estate analysis and decision-making. |
| Keywords: | Digitalisation; LLMs; NLP; real estate |
| JEL: | R3 |
| Date: | 2024–01–01 |
| URL: | https://d.repec.org/n?u=RePEc:arz:wpaper:eres2024-036 |
| By: | Baptiste Lefort; Eric Benhamou; Jean-Jacques Ohana; David Saltiel; Beatrice Guez |
| Abstract: | 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. |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2409.11408 |
| By: | Nigar Karimova |
| Abstract: | The research investigates how the application of a machine-learning random forest model improves the accuracy and precision of a Delphi model. The context of the research is Azerbaijani SMEs and the data for the study has been obtained from a financial institution which had gathered it from the enterprises (as there is no public data on local SMEs, it was not practical to verify the data independently). The research used accuracy, precision, recall and F-1 scores for both models to compare them and run the algorithms in Python. The findings showed that accuracy, precision, recall and F- 1 all improve considerably (from 0.69 to 0.83, from 0.65 to 0.81, from 0.56 to 0.77 and from 0.58 to 0.79, respectively). The implications are that by applying AI models in credit risk modeling, financial institutions can improve the accuracy of identifying potential defaulters which would reduce their credit risk. In addition, an unfair rejection of credit access for SMEs would also go down having a significant contribution to an economic growth in the economy. Finally, such ethical issues as transparency of algorithms and biases in historical data should be taken on board while making decisions based on AI algorithms in order to reduce mechanical dependence on algorithms that cannot be justified in practice. |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2410.05330 |
| By: | Sophia Bodensteiner; Lukas Lautenschlaeger; Wolfgang Schäfers |
| Abstract: | Twitter is established as a major platform for sharing information and opinions online. This research explores the impact of the sentiment expressed on Twitter on the indirect U.S. real estate market, particularly focusing on financial metrics such as returns and volatility. It analyzes how Twitter sentiment correlates with the overall indirect market and additionally focuses on the corporate level, investigating if general findings are also applicable on an individual company basis. Given by the nature of Twitter messages, comprehensive natural language processing is applied to clean and identify relevant posts and to provide the foundation for extracting the sentiment. The complex linguistic features of the given informal language are handled by using three different approaches for classification including traditional and advanced machine-learning methods. Preliminary results suggest that social media sentiment holds predictive value for both market trends and corporate-level changes. Moreover, they indicate towards changing dynamics in the impact of market sentiment on performance metrics during a crisis, exemplified by the 2020 COVID-19 pandemic. This research additionally highlights the effectiveness of classical dictionary-based approaches for sentiment analysis but also shows the capabilities of more sophisticated classifiers. |
| Keywords: | Machine Learning; REITs; Sentiment Analysis; Twitter |
| JEL: | R3 |
| Date: | 2024–01–01 |
| URL: | https://d.repec.org/n?u=RePEc:arz:wpaper:eres2024-090 |
| By: | Miroslav Pánik; Andrej Adamušin |
| Abstract: | The current turbulent development in the real estate market, not only in Slovakia but throughout Europe, is caused by several crises that have a global impact and partially resemble the situation in the market 15 years ago. One factor that negatively affects the real estate market in Slovakia is the lack of residential real estate. In the long term, the demand for housing significantly exceeds the supply of apartments.The article analyzes the prices of real estate intended for housing in Slovakia. It defines economic, demographic and social factors that have a significant impact on the development of real estate prices. The methodology of statistical data collection differs considerably in the developed countries of the EU. Národná banka Slovenska has been publishing quarterly data on real estate prices since 2005. Development modeling is possible, for example, using correlation and regression analysis, which is well known. In the presented article, the forecast of housing prices will be realized with the help of artificial intelligence - neural networks. |
| Keywords: | Artificial Intelligence; Forecasting; Neural Networks; residential real estate prices |
| JEL: | R3 |
| Date: | 2024–01–01 |
| URL: | https://d.repec.org/n?u=RePEc:arz:wpaper:eres2024-091 |
| By: | Siyu Bie (City University of Hong Kong); Francis X. Diebold (University of Pennsylvania); Jingyu He (City University of Hong Kong); Junye Li (Fudan University) |
| Abstract: | We explore tree-based macroeconomic regime-switching in the context of the dynamic Nelson-Siegel (DNS) yield-curve model. In particular, we customize the treegrowing algorithm to partition macroeconomic variables based on the DNS model’s marginal likelihood, thereby identifying regime-shifting patterns in the yield curve. Compared to traditional Markov-switching models, our model offers clear economic interpretation via macroeconomic linkages and ensures computational simplicity. In an empirical application to U.S. Treasury bond yields, we find (1) important yield curve regime switching, and (2) evidence that macroeconomic variables have predictive power for the yield curve when the short rate is high, but not in other regimes, thereby refining the notion of yield curve “macro-spanning”. |
| Keywords: | Decision Tree; Macro-Finance; Term Structure; Regime Switching; Dynamic Nelson-Siegel Model; Bayesian Estimation |
| JEL: | C11 E43 G12 |
| Date: | 2024–10–08 |
| URL: | https://d.repec.org/n?u=RePEc:pen:papers:24-028 |
| By: | Siqi Huang; Anupam Nanda; Eero Valtonen |
| Abstract: | Research on ESG issues in real estate has indicated a complex link between sustainable real estate and financial success. A particular challenge in exploring the link comes from the lack of robust measures of various aspects of real estate sustainability, partly due to the limited availability of granular data. Presently, the assessment of firm-level real estate sustainability mainly uses ESG or CSR ratings from rating agencies, which rely heavily on self-disclosed ESG data from real estate companies. These third-party ESG/CSR ratings come with several limitations, including overlooking the potential impact of disclosure language patterns and sentiment bias on ESG ratings. In this study, we focus on the words around ESG actions and promises from official disclosures and credible news sources and offer a comprehensive evaluation of real estate companies' sustainability performance. We focus on 65 US REITs over a 15-year time period. Our preliminary analysis, based on computational linguistic techniques, highlights various salient features of what companies disclose in the public domain. |
| Keywords: | ESG Discourse; Real Estate Sustainability |
| JEL: | R3 |
| Date: | 2024–01–01 |
| URL: | https://d.repec.org/n?u=RePEc:arz:wpaper:eres2024-076 |