|
on Big Data |
| By: | Imad Talhartit (Université Hassan 1er [Settat], Ecole Nationale de Commerce et Gestion - Settat, Laboratory of Finance, Audit and Organizational Governance Research); Sanae Ait Jillali (Université Hassan 1er [Settat], Ecole Nationale de Commerce et Gestion - Settat, Laboratory of Finance, Audit and Organizational Governance Research); Mounime El Kabbouri (Université Hassan 1er [Settat], Ecole Nationale de Commerce et Gestion - Settat, Laboratory of Finance, Audit and Organizational Governance Research) |
| Abstract: | In today's data-driven economy, predicting stock market behavior has become a key focus for both finance professionals and academics. Traditionally reliant on historical and economic data, stock price forecasting is now being enhanced by AI technologies, especially Deep Learning and Natural Language Processing (NLP), which allow the integration of qualitative data like news sentiment and investor opinions. Deep Learning uses multi-layered neural networks to analyze complex patterns, while NLP enables machines to interpret human language, making it useful for extracting sentiment from media sources. Though most research has focused on developed markets, emerging economies like Morocco offer a unique context due to their evolving financial systems and data limitations. This study takes a theoretical and exploratory approach, aiming to conceptually examine how macroeconomic indicators and sentiment analysis can be integrated using deep learning models to enhance stock price prediction in Morocco. Rather than building a model, the paper reviews literature, evaluates data sources, and identifies key challenges and opportunities. Ultimately, the study aims to bridge AI techniques with financial theory in an emerging market setting, providing a foundation for future empirical research and interdisciplinary collaboration. |
| Keywords: | Stock Price Prediction, Deep Learning, Natural Language Processing (NLP), Sentiment Analysis, Macroeconomic Indicators, Emerging Markets, Moroccan Financial Market |
| Date: | 2025–05 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05094029 |
| By: | Shunyao Wang; Ming Cheng; Christina Dan Wang |
| Abstract: | Stochastic Discount Factor (SDF) models provide a unified framework for asset pricing and risk assessment, yet traditional formulations struggle to incorporate unstructured textual information. We introduce NewsNet-SDF, a novel deep learning framework that seamlessly integrates pretrained language model embeddings with financial time series through adversarial networks. Our multimodal architecture processes financial news using GTE-multilingual models, extracts temporal patterns from macroeconomic data via LSTM networks, and normalizes firm characteristics, fusing these heterogeneous information sources through an innovative adversarial training mechanism. Our dataset encompasses approximately 2.5 million news articles and 10, 000 unique securities, addressing the computational challenges of processing and aligning text data with financial time series. Empirical evaluations on U.S. equity data (1980-2022) demonstrate NewsNet-SDF substantially outperforms alternatives with a Sharpe ratio of 2.80. The model shows a 471% improvement over CAPM, over 200% improvement versus traditional SDF implementations, and a 74% reduction in pricing errors compared to the Fama-French five-factor model. In comprehensive comparisons, our deep learning approach consistently outperforms traditional, modern, and other neural asset pricing models across all key metrics. Ablation studies confirm that text embeddings contribute significantly more to model performance than macroeconomic features, with news-derived principal components ranking among the most influential determinants of SDF dynamics. These results validate the effectiveness of our multimodal deep learning approach in integrating unstructured text with traditional financial data for more accurate asset pricing, providing new insights for digital intelligent decision-making in financial technology. |
| Date: | 2025–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2505.06864 |
| By: | Wishnu Badrawani |
| Abstract: | This paper evaluates the performance of prominent machine learning (ML) algorithms in predicting Indonesia's inflation using the payment system, capital market, and macroeconomic data. We compare the forecasting performance of each ML model, namely shrinkage regression, ensemble learning, and super vector regression, to that of the univariate time series ARIMA and SARIMA models. We examine various out-of-bag sample periods in each ML model to determine the appropriate data-splitting ratios for the regression case study. This study indicates that all ML models produced lower RMSEs and reduced average forecast errors by 45.16 percent relative to the ARIMA benchmark, with the Extreme Gradient Boosting model outperforming other ML models and the benchmark. Using the Shapley value, we discovered that numerous payment system variables significantly predict inflation. We explore the ML forecast using local Shapley decomposition and show the relationship between the explanatory variables and inflation for interpretation. The interpretation of the ML forecast highlights some significant findings and offers insightful recommendations, enhancing previous economic research that uses a more established econometric method. Our findings advocate ML models as supplementary tools for the central bank to predict inflation and support monetary policy. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.10369 |
| By: | YANG, ZHANG (Department of Finance and Business Economics, Faculty of Business Administration / Asia-Pacific Academy of Economics and Management, University of Macau); JIANXIONG LIN (QIFU Technology, China); YIHE QIAN (Department of Finance and Business Economics, Faculty of Business Administration, University of Macau); LIANJIE SHU (Faculty of Business Administration , University of Macau) |
| Abstract: | MachiAs a key enabler of poverty alleviation and equitable growth, financial inclusion aims to expand access to credit and financial services for underserved individuals and small businesses. However, the elevated default risk and data scarcity in inclusive lending present major challenges to traditional credit assessment tools. This study evaluates whether machine learning (ML) techniques can improve default prediction for small-business loans, thereby enhancing the effectiveness and fairness of credit allocation. Using proprietary loan-level data from a city commercial bank in China, we compare eight classification models—Logistic Regression, Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Tree, Random Forest, XGBoost, and LightGBM—under three sampling strategies to address class imbalance. Our findings reveal that undersampling significantly enhances model performance, and tree-based ML models, particularly XGBoost and Decision Tree, outperform traditional classifiers. Feature importance and misclassification analyses suggest that documentation completeness, demographic traits, and credit utilization are critical predictors of default. By combining robust empirical validation with model interpretability, this study contributes to the growing literature at the intersection of machine learning, credit risk, and financial development. Our findings offer actionable insights for policymakers, financial institutions, and data scientists working to build fairer and more effective credit systems in emerging markets. |
| Keywords: | machine learning, financial inclusion, small business, China, credit risk assessment |
| JEL: | G21 G32 C53 O16 |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:boa:wpaper:202532 |
| By: | Schulthess, Urs; Fonteyne, Simon; Gardeazabal Monsalve, Andrea |
| Abstract: | This study aimed to assess whether radar (Sentinel-1) and optical (Sentinel-2) satellite data could detect residue management practices and differentiate between conventional, minimal, and no tillage fields in Guanajuato, Mexico. The study used in-situ data collected by the CIMMYT-led MasAgro Guanajuato project, which tracks land preparation and crop management. Various tillage and residue indices were tested, including NDSVI, NDTI, and NDI5, based on Sentinel-2 bands. The conclusion suggests that most successful remote sensing applications for tillage detection and residue management rely on survey data. These data can then be used to train machine learning based algorithms. |
| Keywords: | remote sensing; crop residue management; conservation agriculture; tillage; Mexico; Latin America |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:fpr:cgiarp:173025 |
| By: | Haoyuan Wang; Chen Liu; Minh-Ngoc Tran; Chao Wang |
| Abstract: | This paper introduces a novel multivariate volatility modeling framework, named Long Short-Term Memory enhanced BEKK (LSTM-BEKK), that integrates deep learning into multivariate GARCH processes. By combining the flexibility of recurrent neural networks with the econometric structure of BEKK models, our approach is designed to better capture nonlinear, dynamic, and high-dimensional dependence structures in financial return data. The proposed model addresses key limitations of traditional multivariate GARCH-based methods, particularly in capturing persistent volatility clustering and asymmetric co-movement across assets. Leveraging the data-driven nature of LSTMs, the framework adapts effectively to time-varying market conditions, offering improved robustness and forecasting performance. Empirical results across multiple equity markets confirm that the LSTM-BEKK model achieves superior performance in terms of out-of-sample portfolio risk forecast, while maintaining the interpretability from the BEKK models. These findings highlight the potential of hybrid econometric-deep learning models in advancing financial risk management and multivariate volatility forecasting. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.02796 |
| By: | J\k{e}drzej Maskiewicz; Pawe{\l} Sakowski |
| Abstract: | The paper explores the use of Deep Reinforcement Learning (DRL) in stock market trading, focusing on two algorithms: Double Deep Q-Network (DDQN) and Proximal Policy Optimization (PPO) and compares them with Buy and Hold benchmark. It evaluates these algorithms across three currency pairs, the S&P 500 index and Bitcoin, on the daily data in the period of 2019-2023. The results demonstrate DRL's effectiveness in trading and its ability to manage risk by strategically avoiding trades in unfavorable conditions, providing a substantial edge over classical approaches, based on supervised learning in terms of risk-adjusted returns. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.04658 |
| By: | Daniel Souza; Aldo Geuna; Jeff Rodríguez |
| Abstract: | We investigate the emergence of Deep Learning as a technoscientific field, emphasizing the role of open labeled datasets. Through qualitative and quantitative analyses, we evaluate the role of datasets like CIFAR-10 in advancing computer vision and object recognition, which are central to the Deep Learning revolution. Our findings highlight CIFAR-10’s crucial role and enduring influence on the field, as well as its importance in teaching ML techniques. Results also indicate that dataset characteristics such as size, number of instances, and number of categories, were key factors. Econometric analysis confirms that CIFAR-10, a small-but- sufficiently-large open dataset, played a significant and lasting role in technological advancements and had a major function in the development of the early scientific literature as shown by citation metrics. |
| Keywords: | Artificial Intelligence; Deep Learning; Emergence of technosciences; Open science; Open Labeled Datasets |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:cca:wpaper:738 |
| By: | Feliks Ba\'nka; Jaros{\l}aw A. Chudziak |
| Abstract: | Accurate option pricing is essential for effective trading and risk management in financial markets, yet it remains challenging due to market volatility and the limitations of traditional models like Black-Scholes. In this paper, we investigate the application of the Informer neural network for option pricing, leveraging its ability to capture long-term dependencies and dynamically adjust to market fluctuations. This research contributes to the field of financial forecasting by introducing Informer's efficient architecture to enhance prediction accuracy and provide a more adaptable and resilient framework compared to existing methods. Our results demonstrate that Informer outperforms traditional approaches in option pricing, advancing the capabilities of data-driven financial forecasting in this domain. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.05565 |
| By: | Konstantin Boss; Luigi Longo; Luca Onorante |
| Abstract: | Using a state-of-the-art large language model, we extract forward-looking and context-sensitive signals related to inflation and unemployment in the euro area from millions of Reddit submissions and comments. We develop daily indicators that incorporate, in addition to posts, the social interaction among users. Our empirical results show consistent gains in out-of-sample nowcasting accuracy relative to daily newspaper sentiment and financial variables, especially in unusual times such as the (post-)COVID-19 period. We conclude that the application of AI tools to the analysis of social media, specifically Reddit, provides useful signals about inflation and unemployment in Europe at daily frequency and constitutes a useful addition to the toolkit available to economic forecasters and nowcasters. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.10546 |
| By: | Philipp Koch; Viktor Stojkoski; C\'esar A. Hidalgo |
| Abstract: | Can we use data on the biographies of historical figures to estimate the GDP per capita of countries and regions? Here we introduce a machine learning method to estimate the GDP per capita of dozens of countries and hundreds of regions in Europe and North America for the past 700 years starting from data on the places of birth, death, and occupations of hundreds of thousands of historical figures. We build an elastic net regression model to perform feature selection and generate out-of-sample estimates that explain 90% of the variance in known historical income levels. We use this model to generate GDP per capita estimates for countries, regions, and time periods for which this data is not available and externally validate our estimates by comparing them with four proxies of economic output: urbanization rates in the past 500 years, body height in the 18th century, wellbeing in 1850, and church building activity in the 14th and 15th century. Additionally, we show our estimates reproduce the well-known reversal of fortune between southwestern and northwestern Europe between 1300 and 1800 and find this is largely driven by countries and regions engaged in Atlantic trade. These findings validate the use of fine-grained biographical data as a method to produce historical GDP per capita estimates. We publish our estimates with confidence intervals together with all collected source data in a comprehensive dataset. |
| Date: | 2025–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2505.09399 |
| By: | Emily Aiken; Anik Ashraf; Joshua E. Blumenstock; Raymond P. Guiteras; Ahmed Mushfiq Mobarak |
| Abstract: | Innovations in big data and algorithms are enabling new approaches to target interventions at scale. We compare the accuracy of three different systems for identifying the poor to receive benefit transfers - proxy means-testing, nominations from community members, and an algorithmic approach using machine learning to predict poverty using mobile phone usage behavior - and study how their cost-effectiveness varies with the scale and scope of the program. We collect mobile phone records from all major telecom operators in Bangladesh and conduct community-based wealth rankings and detailed consumption surveys of 5, 000 households, to select the 22, 000 poorest households for $300 transfers from 106, 000 listed households. While proxy-means testing is most accurate, algorithmic targeting becomes more cost-effective for national-scale programs where large numbers of households have to be screened. We explore the external validity of these insights using survey data and mobile phone records data from Togo, and cross-country information on benefit transfer programs from the World Bank. |
| Keywords: | cash transfers, digital data, impact evaluation |
| JEL: | C55 I32 I38 |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_11928 |
| By: | Emily Aiken (UC San Diego); Anik Ashraf (Ludwig-Maximilians-Universitat Munchen); Joshua E. Blumenstock (UC Berkeley); Raymond P. Guiteras (NC State University); Ahmed Mushfiq Mobarak (Yale University) |
| Abstract: | Innovations in big data and algorithms are enabling new approaches to target interventions at scale. We compare the accuracy of three different systems for identifying the poor to receive benefit transfers Ñ proxy means-testing, nominations from community members, and an algorithmic approach using machine learning to predict poverty using mobile phone usage behavior Ñ and study how their cost-effectiveness varies with the scale and scope of the program. We collect mobile phone records from all major telecom operators in Bangladesh and conduct community-based wealth rankings and detailed consumption surveys of 5, 000 households, to select the 22, 000 poorest households for $300 transfers from 106, 000 listed households. While proxy-means testing is most accurate, algorithmic targeting becomes more cost-effective for national-scale programs where large numbers of households have to be screened. We explore the external validity of these insights using survey data and mobile phone records data from Togo, and cross-country information on benefit transfer programs from the World Bank. |
| Date: | 2025–06–08 |
| URL: | https://d.repec.org/n?u=RePEc:cwl:cwldpp:2443 |
| By: | Issa Sugiura; Takashi Ishida; Taro Makino; Chieko Tazuke; Takanori Nakagawa; Kosuke Nakago; David Ha |
| Abstract: | Financial analysis presents complex challenges that could leverage large language model (LLM) capabilities. However, the scarcity of challenging financial datasets, particularly for Japanese financial data, impedes academic innovation in financial analytics. As LLMs advance, this lack of accessible research resources increasingly hinders their development and evaluation in this specialized domain. To address this gap, we introduce EDINET-Bench, an open-source Japanese financial benchmark designed to evaluate the performance of LLMs on challenging financial tasks including accounting fraud detection, earnings forecasting, and industry prediction. EDINET-Bench is constructed by downloading annual reports from the past 10 years from Japan's Electronic Disclosure for Investors' NETwork (EDINET) and automatically assigning labels corresponding to each evaluation task. Our experiments reveal that even state-of-the-art LLMs struggle, performing only slightly better than logistic regression in binary classification for fraud detection and earnings forecasting. These results highlight significant challenges in applying LLMs to real-world financial applications and underscore the need for domain-specific adaptation. Our dataset, benchmark construction code, and evaluation code is publicly available to facilitate future research in finance with LLMs. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.08762 |
| By: | Haochuan (Kevin); Wang |
| Abstract: | Cryptocurrency price dynamics are driven largely by microstructural supply demand imbalances in the limit order book (LOB), yet the highly noisy nature of LOB data complicates the signal extraction process. Prior research has demonstrated that deep-learning architectures can yield promising predictive performance on pre-processed equity and futures LOB data, but they often treat model complexity as an unqualified virtue. In this paper, we aim to examine whether adding extra hidden layers or parameters to "blackbox ish" neural networks genuinely enhances short term price forecasting, or if gains are primarily attributable to data preprocessing and feature engineering. We benchmark a spectrum of models from interpretable baselines, logistic regression, XGBoost to deep architectures (DeepLOB, Conv1D+LSTM) on BTC/USDT LOB snapshots sampled at 100 ms to multi second intervals using publicly available Bybit data. We introduce two data filtering pipelines (Kalman, Savitzky Golay) and evaluate both binary (up/down) and ternary (up/flat/down) labeling schemes. Our analysis compares models on out of sample accuracy, latency, and robustness to noise. Results reveal that, with data preprocessing and hyperparameter tuning, simpler models can match and even exceed the performance of more complex networks, offering faster inference and greater interpretability. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.05764 |
| By: | Duosi Zheng; Hanzhong Guo; Yanchu Liu; Wei Huang |
| Abstract: | Recognizing the importance of jump risk in option pricing, we propose a neural jump stochastic differential equation model in this paper, which integrates neural networks as parameter estimators in the conventional jump diffusion model. To overcome the problem that the backpropagation algorithm is not compatible with the jump process, we use the Gumbel-Softmax method to make the jump parameter gradient learnable. We examine the proposed model using both simulated data and S&P 500 index options. The findings demonstrate that the incorporation of neural jump components substantially improves the accuracy of pricing compared to existing benchmark models. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.05137 |
| By: | Undral Byambadalai; Tomu Hirata; Tatsushi Oka; Shota Yasui |
| Abstract: | This paper focuses on the estimation of distributional treatment effects in randomized experiments that use covariate-adaptive randomization (CAR). These include designs such as Efron's biased-coin design and stratified block randomization, where participants are first grouped into strata based on baseline covariates and assigned treatments within each stratum to ensure balance across groups. In practice, datasets often contain additional covariates beyond the strata indicators. We propose a flexible distribution regression framework that leverages off-the-shelf machine learning methods to incorporate these additional covariates, enhancing the precision of distributional treatment effect estimates. We establish the asymptotic distribution of the proposed estimator and introduce a valid inference procedure. Furthermore, we derive the semiparametric efficiency bound for distributional treatment effects under CAR and demonstrate that our regression-adjusted estimator attains this bound. Simulation studies and empirical analyses of microcredit programs highlight the practical advantages of our method. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.05945 |
| By: | Fedele, Alessandro (Free University of Bozen/Bolzano); Tonin, Mirco (Free University of Bozen/Bolzano); Wiesen, Daniel (University of Cologne) |
| Abstract: | The health sector requires skilled, altruistic, and motivated individuals to perform complex tasks for which ex-post incentives may prove ineffective. Understanding the determinants of self-selection into health professions is therefore critical. We investigate this issue relying on data from surveys and incentivized dictator games. We compare applicants to medical and healthcare schools in Italy and Austria with non-applicants from the same regions and age cohorts. Drawing on a wide range of individual characteristics, we employ machine learning techniques for variable selection. Our findings show that higher cognitive ability, greater altruism, and the personality trait of conscientiousness are positively associated with the likelihood of applying to medical or nursing school, while neuroticism is negatively associated. Additionally, individuals with a strong identification with societal goals and those with parents working as doctors are more likely to pursue medical education. These results provide evidence of capable, altruistic, and motivated individuals self-selecting into the health sector, a necessary condition for building a high-quality healthcare workforce. |
| Keywords: | personality traits, cognitive ability, altruism, health professions, self-selection, machine learning |
| JEL: | I1 J24 J4 |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp17941 |
| By: | Hasan Fallahgoul |
| Abstract: | Recent advances in machine learning have shown promising results for financial prediction using large, over-parameterized models. This paper provides theoretical foundations and empirical validation for understanding when and how these methods achieve predictive success. I examine three key aspects of high-dimensional learning in finance. First, I prove that within-sample standardization in Random Fourier Features implementations fundamentally alters the underlying Gaussian kernel approximation, replacing shift-invariant kernels with training-set dependent alternatives. Second, I derive sample complexity bounds showing when reliable learning becomes information-theoretically impossible under weak signal-to-noise ratios typical in finance. Third, VC-dimension analysis reveals that ridgeless regression's effective complexity is bounded by sample size rather than nominal feature dimension. Comprehensive numerical validation confirms these theoretical predictions, revealing systematic breakdown of claimed theoretical properties across realistic parameter ranges. These results show that when sample size is small and features are high-dimensional, observed predictive success is necessarily driven by low-complexity artifacts, not genuine high-dimensional learning. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.03780 |
| By: | Cao, Jingcun; Li, Xiaolin; Zhang, Lingling |
| Abstract: | Firms increasingly use a combination of image and text description when displaying products and engaging consumers. Existing research has examined consumers’ response to text and image stimuli separately but has yet to systematically consider how the semantic relationship between image and text impacts consumer choice. In this research, we conduct a series of multimethod empirical studies to examine the congruence between image- and text-based product representation. First, we propose a deep-learning approach to measure image-text congruence by building a state-of-the-art two-branch neural network model based on wide residual networks and bidirectional encoder representations from transformers. Next, we apply our method to data from an online reading platform and discover a U-shaped effect of image-text congruence: Consumers’ preference toward a product is higher when the congruence between the image and text representation is either high or low than when the congruence is at the medium level. We then conduct experiments to establish the causal effect of this finding and explore the underlying mechanisms. We further explore the generalizability of the proposed deep-learning model and our substantive finding in two additional settings. Our research contributes to the literature on consumer information processing and generates managerial implications for practitioners on how to strategically pair images and text on digital platforms. |
| JEL: | L81 |
| Date: | 2025–05–09 |
| URL: | https://d.repec.org/n?u=RePEc:ehl:lserod:128215 |
| By: | Harold D. Chiang; Jack Collison; Lorenzo Magnolfi; Christopher Sullivan |
| Abstract: | This paper develops a flexible approach to predict the price effects of horizontal mergers using ML/AI methods. While standard merger simulation techniques rely on restrictive assumptions about firm conduct, we propose a data-driven framework that relaxes these constraints when rich market data are available. We develop and identify a flexible nonparametric model of supply that nests a broad range of conduct models and cost functions. To overcome the curse of dimensionality, we adapt the Variational Method of Moments (VMM) (Bennett and Kallus, 2023) to estimate the model, allowing for various forms of strategic interaction. Monte Carlo simulations show that our method significantly outperforms an array of misspecified models and rivals the performance of the true model, both in predictive performance and counterfactual merger simulations. As a way to interpret the economics of the estimated function, we simulate pass-through and reveal that the model learns markup and cost functions that imply approximately correct pass-through behavior. Applied to the American Airlines-US Airways merger, our method produces more accurate post-merger price predictions than traditional approaches. The results demonstrate the potential for machine learning techniques to enhance merger analysis while maintaining economic structure. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.05225 |
| By: | Stephane Hess; Sander van Cranenburgh |
| Abstract: | Travel behaviour modellers have an increasingly diverse set of models at their disposal, ranging from traditional econometric structures to models from mathematical psychology and data-driven approaches from machine learning. A key question arises as to how well these different models perform in prediction, especially when considering trips of different characteristics from those used in estimation, i.e. out-of-distribution prediction, and whether better predictions can be obtained by combining insights from the different models. Across two case studies, we show that while data-driven approaches excel in predicting mode choice for trips within the distance bands used in estimation, beyond that range, the picture is fuzzy. To leverage the relative advantages of the different model families and capitalise on the notion that multiple `weak' models can result in more robust models, we put forward the use of a model averaging approach that allocates weights to different model families as a function of the \emph{distance} between the characteristics of the trip for which predictions are made, and those used in model estimation. Overall, we see that the model averaging approach gives larger weight to models with stronger behavioural or econometric underpinnings the more we move outside the interval of trip distances covered in estimation. Across both case studies, we show that our model averaging approach obtains improved performance both on the estimation and validation data, and crucially also when predicting mode choices for trips of distances outside the range used in estimation. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.03693 |
| By: | Hadi Hosseini; Samarth Khanna; Ronak Singh |
| Abstract: | The rise of Large Language Models (LLMs) has driven progress in reasoning tasks -- from program synthesis to scientific hypothesis generation -- yet their ability to handle ranked preferences and structured algorithms in combinatorial domains remains underexplored. We study matching markets, a core framework behind applications like resource allocation and ride-sharing, which require reconciling individual ranked preferences to ensure stable outcomes. We evaluate several state-of-the-art models on a hierarchy of preference-based reasoning tasks -- ranging from stable-matching generation to instability detection, instability resolution, and fine-grained preference queries -- to systematically expose their logical and algorithmic limitations in handling ranked inputs. Surprisingly, even top-performing models with advanced reasoning struggle to resolve instability in large markets, often failing to identify blocking pairs or execute algorithms iteratively. We further show that parameter-efficient fine-tuning (LoRA) significantly improves performance in small markets, but fails to bring about a similar improvement on large instances, suggesting the need for more sophisticated strategies to improve LLMs' reasoning with larger-context inputs. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.04478 |
| By: | Shunxin Yao |
| Abstract: | OC-DeepIV is a neural network model designed for estimating causal effects. It characterizes heterogeneity by adding interaction features and reduces redundancy through orthogonal constraints. The model includes two feature extractors, one for the instrumental variable Z and the other for the covariate X*. The training process is divided into two stages: the first stage uses the mean squared error (MSE) loss function, and the second stage incorporates orthogonal regularization. Experimental results show that this model outperforms DeepIV and DML in terms of accuracy and stability. Future research directions include applying the model to real-world problems and handling scenarios with multiple processing variables. |
| Date: | 2025–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2506.02790 |