Computational Economics http://lists.repec.org/mailman/listinfo/nep-cmp Computational Economics 2026-07-13 Data-Driven Duration Management -- Term Structure Forecasting Using Machine Learning https://d.repec.org/n?u=RePEc:arx:papers:2606.26815&r=&r=cmp This paper compares different methods for forecasting the term structure of U.S. and European zero-coupon government bonds using both traditional econometric and Machine Learning (ML) approaches. We compare classical models (e.g., Dynamic Nelson-Siegel (DNS) and Principal Component Analysis (PCA)) with different Neural Network (NN) architectures, including those inspired by the classical models, on the U.S. Treasury market and bonds issued by the European Central Bank (ECB). To enhance predictive performance, macroeconomic variables are incorporated. The findings for both markets are separately analyzed and compared. To this end, we propose a robust model evaluation framework combining statistical accuracy metrics - such as RMSE, MAE, and directional accuracy - with the economic relevance of a quantitative bond trading strategy. Results show that NNs consistently outperform traditional models in both forecasting accuracy and portfolio performance. For the U.S., the most effective approach is a direct-forecasting NN that incorporates DNS factors to reduce the dimensionality of zero-rate data and an Autoencoder (AE) to extract macroeconomic features, while for Europe, the optimal model is a factor-based NN using PCA-derived zero-rate factors without the integration of macroeconomic variables. Overall, the paper demonstrates how combining traditional modeling approaches with modern ML techniques and evaluation can improve yield curve forecasts and support applications in fixed-income portfolio construction. Tobias Lausser Joao Eduardo Vuolo Rudi Zagst 2026-06 Recurrent Neural Networks for real estate evaluation in the Italian market https://d.repec.org/n?u=RePEc:ven:wpaper:2026:22&r=&r=cmp Accurate evaluation of real estate prices represents a complex task, that holds significant relevance for the majority of the actors involved in a society, as homebuyers/sellers, investors, and policymakers. Over the last decades, in parallel with technological advancements, scholars and practitioners proposed several models and methodologies, spanning from simple hedonic regression models to sophisticated machine learning algorithms, in the attempt to improve the accuracy of price forecasts. Nonetheless, the interest towards the development of more precise and performing methodologies remains high, not only for static valuation but also, more recently, for dynamic approaches. Focusing on dynamic valuation, we turn the attention towards the Recurrent Neural Network (RNN) models: while widely tested for stock prices forecast, RNNs have seen limited to no application in the domain of real estate valuation. In this study, using official data on prices of Italian residential properties, we propose and implement two different architectures for RNN models to perform both price prediction and relevance analysis with respect to the explanatory variables. Differently from other studies, we utilize data for the whole Italian territory, focusing, rather than on individual properties' valuation, on a range of prices for groups of properties identified by house type, geographical position and state of conservation. Specifically, the output variables of our analysis are represented by the minimum and maximum prices that determine this range. The results of our analysis indicate that a Standard Recurrent Neural Network forecasts well the prices in small municipalities (with less than 15 000 inhabitants), while a Double-Input Layer RNN, with an input layer dedicated to static features and a separate input layer for dynamic features, performs better in medium-sized (with a number of inhabitants between 15 000 and 50 000 inhabitants) and large municipalities (with more than 50 000 municipalities). Antonella Basso Marco Corazza Lorenzo Tonon Real Estate valuation, House price forecast, Recurrent Neural Networks, Machine Learning, LSTM Networks 2026 How Well Do LLMs Predict Human Behavior? A Measure of their Pretrained Knowledge* https://d.repec.org/n?u=RePEc:bri:uobdis:26/835&r=&r=cmp Large language models (LLMs) are increasingly used to predict human behavior. We propose a measure for evaluating how much knowledge a pretrained LLM brings to such a prediction: its equivalent sample size, defined as the amount of task-specific data needed to match the predictive accuracy of the LLM. We estimate this measure by comparing the prediction error of a fixed LLM in a given domain to that of flexible machine learning models trained on increasing samples of domain-specific data. We further provide a statistical inference procedure by developing a new asymptotic theory for cross-validated prediction error. Finally, we apply this method to the Panel Study of Income Dynamics. We find that LLMs encode considerable predictive information for some economic variables but much less for others, suggesting that their value as substitutes for domain-specific data differs markedly across settings. Wayne Gao Sukjin Han Annie Liang 2026-01-30 Cascading Multi-Agent Policy Optimization for Demand Forecasting https://d.repec.org/n?u=RePEc:hal:journl:hal-05656779&r=&r=cmp Reliable demand forecasting is crucial for effective supply chain management, where inaccurate forecasts can lead to frequent out-of-stock or overstock situations. While numerous statistical and machine learning methods have been explored for demand forecasting, reinforcement learning approaches, despite their significant potential, remain little known in this domain. In this paper, we propose a multi-agent deep reinforcement learning solution designed to accurately predict demand across multiple stores. We present empirical evidence that demonstrates the effectiveness of our model using a real-world dataset. The results confirm the practicality of our proposed approach and highlight its potential to improve demand forecasting in retail and potentially other forecasting scenarios. Saeed Varasteh Yazdi reinforcement learning, multi-agent systems, demand forecasting 2025-07-31 Artificial Intelligence in Ship Finance: Applications, Opportunities, and a Case Study in AI-Augmented Loan Origination https://d.repec.org/n?u=RePEc:arx:papers:2606.11238&r=&r=cmp Ship finance is a data-intensive and document-heavy segment of asset-based lending, requiring the integration of financial, technical, contractual, and regulatory information from heterogeneous and largely unstructured sources. Increasing environmental regulation and ESG reporting requirements are adding further complexity to underwriting and loan-origination processes. Recent advances in artificial intelligence (AI), particularly large language models (LLMs), create new opportunities for processing and analysing such information. This paper reviews potential applications of AI in ship finance, with a particular focus on LLM-based systems for document comprehension, information extraction, and workflow automation. We present ShipFinance.ai, a modular agentic architecture to support loan application workflows in ship finance. The proposed system combines an LLM-based extraction module, financial analysis components, external maritime data services, and a controlled document-generation module with a chatbot interface to support the preparation of standardized financing applications. The paper discusses the key challenges for using such models in production. We argue that AI-assisted systems can support maritime finance professionals in managing increasingly complex information and reporting requirements. Lasse Dierich Orestis Schinas 2026-05 How to deal with machine learning bias in economic history https://d.repec.org/n?u=RePEc:hes:wpaper:0306&r=&r=cmp Machine learning (ML) has rapidly transformed economic history, lowering costs of digitization, data linkage, and imputation, and making information in historical text usable at scale. This paper offers a practical guide to using these tools well. However, ML tools have also created new problems. Prediction errors are often systematically correlated with covariates of interest, so even highly accurate models can distort and sometimes reverse coefficients, and standard validation cannot detect this. Given that ML tools often perform worse for historical data, this problem is especially severe for the field of economic history. We also identify a solution to this problem. We show that recent debiasing methods can correct such bias for a wide class of applications, using a small, randomly sampled set of expert-coded labels while retaining the efficiency of large-scale prediction. We organize the field with a taxonomy of three ML tasks, survey the literature along it, and indicate where debiasing applies and where validation against proxies remains the only recourse. We close with best-practice guidance on digitization, model choice, and reproducibility. Torben S. D. Johansen Julius Koschnick Christian Vedel Machine learning, Large language models, bias correction, economic history 2026-07 Mining Causality: AI-Assisted Search for Instrumental Variables∗ https://d.repec.org/n?u=RePEc:bri:uobdis:26/833&r=&r=cmp The instrumental variables (IVs) method is a leading empirical strategy for causal inference. Finding IVs is a heuristic and creative process, and justifying its validity — especially exclusion restrictions — is largely rhetorical. We propose using large language models (LLMs) to search for new IVs through narratives and counterfactual reasoning, similar to how a human researcher would. The stark difference, however, is that LLMs can dramatically accelerate this process and explore an extremely large search space. We demonstrate how to construct prompts to search for potentially valid IVs. We contend that multi-step and role-playing prompting strategies are effective for simulating the endogenous decision-making processes of economic agents or social actors and for navigating language models through the realm of real-world scenarios. We apply our method to four well-known examples in economics: returns to schooling, demand and supply, and peer effects. Expert surveys reveal that some of the discovered IVs in each domain appear both novel and likely valid. Sukjin Han 2026-01-30 Statistical Foundations of LLM-based A/B Testing: A Surrogacy Framework for Human Causal Inference https://d.repec.org/n?u=RePEc:arx:papers:2606.17165&r=&r=cmp Organizations and researchers show increasing interest in using large language models (LLMs) in place of human participants in A/B tests, in the hope of experimenting faster and at lower cost. We study when a treatment effect estimated on LLM outcomes recovers the effect that would have been measured on the human population of interest. Distributional equivalence between LLM and human outcomes would make any standard estimator valid but is unrealistic. We therefore develop a statistical framework that adapts surrogate endpoint theory to LLMs. The framework shows that calibrating LLM outcomes to human outcomes identifies the average treatment effect under surrogacy and comparability conditions that are jointly weaker than distributional equivalence. When these conditions fail, the effect of interest is only partially identified, and we provide diagnostics that can falsify surrogacy on historical experiments together with a bound on the worst-case bias from limited overlap. We further show that the stochasticity inherent to LLMs introduces both bias and variance, but using an average of multiple draws as the surrogate mitigates both. We illustrate the methods and theory in simulations and an application to A/B tests on Upworthy headlines. A central takeaway from our work is that the validity of LLM outcomes as surrogates can only be falsified for past treatments and never verified for new ones, so human experiments remain indispensable for novel interventions. We discuss the role of LLM choice, prompting, and temperature as design variables, and how to size human experiments for validation. Joel Persson M{\aa}rten Schultzberg Sebastian Ankargren 2026-06 Model Validation of Agentic AI Systems: A POMDP-Based Framework for Belief-State, Forecast, and Policy Validation https://d.repec.org/n?u=RePEc:arx:papers:2606.17383&r=&r=cmp Agentic artificial intelligence systems introduce a new class of model risk. Unlike traditional predictive models, autonomous agents continuously acquire information, form beliefs regarding latent states of the environment, generate forecasts, select actions, and adapt their behavior over time. Existing validation methodologies focus primarily on predictive accuracy and therefore provide limited insight into the quality of the underlying decision process. This paper proposes a model validation framework for agentic AI based on Partially Observable Markov Decision Processes (POMDPs). The framework decomposes autonomous decision making into information, beliefs, forecasts, actions, and utility, allowing each component to be validated independently. Large language models (LLMs) are formalized as approximate Bayesian filtering operators, and a model-risk taxonomy is developed encompassing state-space, filtering, forecast, policy, utility-specification, and parameter risks. The model risk validation methodology is demonstrated through a portfolio-management case study in which an agent infers latent market regimes from market and macroeconomic information, generates belief-conditioned forecasts, and constructs portfolios using a Black--Litterman framework. Empirical validation combines performance analysis, belief calibration diagnostics, coverage tests, ablation studies, and parameter-sensitivity analysis. The results indicate that latent-state inference contributes independently to decision quality and that the principal conclusions remain robust across a broad range of parameter values. The principal contribution of the paper is a practical framework for extending established model risk management concepts to autonomous AI systems and providing a rigorous foundation for their validation, governance, and monitoring. Matthew Francis Dixon 2026-06 Heterogeneous Graph Neural Networks for Product Recommendation on Transactional Retail Data https://d.repec.org/n?u=RePEc:hal:journl:hal-05535428&r=&r=cmp Personalized product recommendation is crucial for enhancing user experience and driving sales in e-commerce, yet effectively leveraging sparse, implicit feedback from transactional data remains challenging. This paper investigates the application of heterogeneous Graph Neural Networks (GNNs) for product recommendation on the widely used Online Retail dataset. We frame the task as link prediction between customers and products, constructing a heterogeneous graph from processed transactional records. We employ a GNN model based on GraphSAGE, adapted for heterogeneity, using learnable embeddings derived solely from the interaction structure. Our experiments demonstrate the model's effectiveness, achieving a ROC AUC of 0.853 and an F1-Score of 0.750 (with 0.937 Recall) on the held-out test set using an optimized configuration (negative sampling ratio 1.0, classification threshold -0.5). We analyze the significant impact of the negative sampling ratio during training on the final precision-recall trade-off, highlighting the importance of aligning training parameters with desired recommendation goals. Our findings confirm the viability of heterogeneous GNNs for modeling implicit feedback and providing effective recommendations in a retail context. This paper is organized as follows: Section I introduces the problem and our approach. Section II reviews related work in recommendation systems and graph-based methods. Section III details our methodology, including dataset description, data preparation, graph construction, and the GNN model architecture. Section IV describes the experimental setup and evaluation metrics. Section V presents the quantitative results, including the analysis of parameters like negative sampling. Section VI discusses the findings and limitations. Finally, Section VII concludes the paper. Karima Belmabrouk Latifa Dekhici Imad Eddine Khiloun Christoph Bergmeir Recommendation Systems, Graph Neural Networks (GNNs), Link Prediction, Implicit Feedback, Heterogeneous Graphs, Online Retail, E-commerce, GraphSAGE, Node Embeddings, Recommendation Systems Graph Neural Networks (GNNs) Link Prediction Implicit Feedback Heterogeneous Graphs Online Retail E-commerce GraphSAGE Node Embeddings 2025-07-23 Place Marketing or Civic Information? Classifying Municipal Tweets using Machine Learning https://d.repec.org/n?u=RePEc:hhs:iuiwop:1563&r=&r=cmp How do municipalities use social media? For place marketing, civic information, or dialogue with citizens? We address this question by classifying 35, 930 tweets from 15 municipal Twitter accounts in the Skåne region of Sweden, using a machine learning algorithm trained on manually annotated data. Tweets are classified along two dimensions: whether they contain place marketing content and whether they contain civic information, yielding four categories. Our findings show that place marketing content declined substantially over the period studied, from nearly 40 percent of tweets in 2009 to just above 20 percent in 2018. In contrast, civic information and direct dialogue with citizens increased. These results suggest that the rise of social media has not trapped municipalities in a zero-sum place marketing arms race. Instead, municipalities appeared to be using Twitter increasingly as a channel for civic communication, with implications for how scholars and practitioners understand the evolving role of social media in place branding. Bergh, Andreas Anzén Ekman, Christina Moricz, Sara Place marketing; Place branding; Social media; Municipalities; Machine learning; Twitter; Sweden 2026-06-22 Pretrained Time-Series Foundation Models for Financial Return Forecasting https://d.repec.org/n?u=RePEc:arx:papers:2606.27100&r=&r=cmp Financial return forecasting is a difficult test case for time-series foundation models (TSFMs) due to low signal-to-noise ratios, structural breaks, heavy tails, and weak persistence. This paper benchmarks pretrained TSFMs against train-from-scratch neural baselines in a deliberately conservative financial setting. We evaluate TimeGPT/TimeGPT-LH, TimesFM-2.5, Moirai-2.0, Chronos, and Chronos-2 against NBEATS, NHITS, PatchTST, iTransformer, and KAN on five liquid U.S. equities (AAPL, AMZN, GOOG, JPM, META) using linear and log returns. Models are compared under an equalized context budget, a rolling-origin protocol, and against random-walk benchmarks. We provide a theoretical framing of pretraining as an inductive prior, linking PAC-Bayes transfer intuition, information-theoretic predictability limits, and attention geometry. This clarifies why strong model rankings need not imply economically meaningful predictability in noisy markets. Pragmatically, pretrained TSFMs dominate the ranking distribution, accounting for 8 of 10 task-level wins. Moirai-2.0 and TimesFM-2.5 achieve the strongest average ranks, leading tasks for AAPL, JPM, GOOG, and AMZN, while Chronos wins the remaining AMZN task. However, the iTransformer baseline wins both META tasks, showing local supervised learning can still outperform generic pretraining for specific assets. Crucially, gains over the random-walk benchmark are small and sparse. A one-sided Diebold-Mariano test rejects equal or inferior predictive accuracy only for Chronos on AMZN and Moirai-2.0 on GOOG. We conclude that TSFMs serve as useful practical priors that reduce model-development costs in low-data financial forecasting, but are not universal engines for statistically reliable alpha generation in realistic empirical deployment. Miquel Noguer I Alonso Rodolfo Pereira Franklin 2026-06