Search Results (98)

This study (N = 249) examines the influence of cognitive reflection on rational decision making in tasks that require the—potentially rapid—integration of multiple pieces of information but are not designed such that intuitive (System 1) responses mislead people. Cognitive reflection was measured using the Cognitive Reflection Test (CRT). Choice behavior was analyzed in 250 probabilistic inference tasks and 16 risky-decision tasks completed by each participant. In both tasks, individuals with higher CRT scores made more rational choices. This superior performance was not attributable to qualitative differences in cognitive processes. For individuals low and high in cognitive reflection, the same Parallel Constraint Satisfaction Model best explained their choice behavior. High-reflective individuals appeared to use the same coherence-based processes more efficiently and consistently. The absence of qualitative process differences across individuals varying in their tendency to engage in deliberate processing supports an integrative account of dual-process models, particularly those grounded in interactive activation frameworks. Full article

Surface faulting poses an earthquake-related hazard with direct consequences for infrastructure and high-risk facilities. Probabilistic Fault Displacement Hazard Analysis (PFDHA) is widely used to estimate the annual frequency of exceedance (AFOE) of specific displacement values at sites on or near active faults. This approach requires numerous input parameters related to fault characterization and coseismic displacement distribution, yet few studies have examined how these parameter choices affect hazard results. Thus, we conduct an analysis following a One-At-a-Time (OAT) strategy, in which a single parameter is varied with respect to three kinematic-specific baselines. We explored the PFDHA outputs obtained allied to the broadly adopted regression models and scaling laws available in the literature up to 2023. We compared the hazard curves obtained for principal faulting from each calculation to a baseline parametrization, and we computed the percentage difference in AFOE, given a displacement amount, with respect to such a baseline. We obtained values in the interval −100% to +200%, computed within the displacement interval adopted for the hazard calculation, attesting that empirical regressions contribute significantly to hazard curve variations. Our sensitivity study could inform operative choices by practitioners and provides insights for optimizing data acquisition efforts in fault displacement hazard assessments. Full article

We investigate how causal DAG learning algorithms and structural assumptions influence counterfactual decision safety. Four structure learning regimes are compared: expert-guided edge-constrained HC+BIC, unconstrained HC+BIC, MMPC+HC+BIC, and the PC-Stable algorithm. Evaluation is conducted using a leave-one-state-out protocol over a fully enumerated state–action space in a controlled offline autonomous driving setting. The environment is characterized by seven Boolean state variables and six actions, allowing us to disentangle the effects of learning strategies on counterfactual decisions. All models are implemented as probabilistic logic twin networks (PLTNs), with additional sensitivity analysis across parameter configurations. The learning regimes produce markedly different counterfactual decisions. Edge-constrained HC+BIC recommends a diverse set of safe actions, while unconstrained HC+BIC yields fewer but consistently safe alternatives. MMPC+HC+BIC frequently fails to identify safe actions, often associated with weak connectivity of the outcome variable. PC-Stable produces varied recommendations but may include unsafe actions, which is linked to incorrect edge orientations between actions and outcomes. These findings show that structure learning choices and prior knowledge influence counterfactual decisions through the learned structure, affecting the identification of safe alternatives in safety-critical applications. Full article

This study investigates the effectiveness of Bayesian probabilistic methods for stock price forecasting on the Johannesburg Stock Exchange by implementing and comparing Gaussian process regression (GPR), Bayesian long short-term memory (Bayesian LSTM), and Bayesian neural networks (BNNs). Using daily open, high, low, close, and volume (OHLCV) data and engineered technical indicators for FirstRand and Discovery from January 2005 to June 2025 (5187 observations), models were trained and evaluated with the mean absolute error (MAE), root mean squared error (RMSE), and mean squared error (MSE). The GPR produced reliable, well-calibrated intervals in relatively stable regimes, but its performance degraded on the more volatile Discovery series. Bayesian LSTM delivered conservative uncertainty estimates with wide predictive intervals but showed the largest point forecast errors. The BNNs achieved the best balance between accuracy and uncertainty quantification, producing the lowest errors for FirstRand and competitive performance for Discovery. Comparative analysis indicates that BNNs are most suitable when point accuracy and calibrated uncertainty are both priorities, GPR is valuable for smaller or more stable data regimes, and Bayesian LSTM is preferable where conservative, risk-conscious intervals are required. This study highlights the practical value of embedding uncertainty into financial forecasts and recommends matching Bayesian model choice to market volatility, data availability, and decision maker risk appetite. Full article

The exponential growth of genomic and proteomic data has made computational protein–protein interaction (PPI) prediction indispensable, driving the need for a comprehensive and method-aware evaluation of supervised learning approaches. PPIs are fundamental to understanding cellular processes and disease mechanisms, yet experimental identification remains slow, costly, and difficult to scale. This survey systematically investigates ten supervised learning models—Extreme Learning Machine (ELM), Convolutional Neural Networks (CNNs), Graph Neural Networks (GNNs), Deep Neural Networks (DNNs), Naïve Bayes, Probabilistic Decision Tree, Support Vector Machine (SVM), Least Squares SVM (LS-SVM), K-Nearest Neighbor (KNN), and Weighted K-Nearest Neighbor (WKNN)—through a tri-layered framework that integrates Comparative Quantitative Analysis, Comparative Observational Analysis, and Experimental Evaluations. Beyond conventional accuracy summaries, this work provides critical commentary tied to real-world use, analyzing where techniques succeed or fail in practice—for instance, when instance-based methods bottleneck during inference, when kernel choices influence SVM variance, or when deep architectures trade accuracy for computational cost. The survey also offers concrete deployment guidance, such as calibration insights for WKNN versus KNN under varying feature noise or dataset curation quality, delivering operational perspectives that typical surveys omit. Comparative Quantitative Analysis consolidates metrics such as accuracy, F1-score, and computational time from the existing literature, while Comparative Observational Analysis evaluates interpretability, scalability, dataset suitability, and efficiency. Complementing these, Experimental Evaluations conducted by the authors empirically validate model performance on benchmark datasets. Together, these layers provide a unified and evidence-backed perspective on algorithmic strengths, weaknesses, and practical applicability. Findings show that GNNs and DNNs achieve the highest predictive accuracy due to their ability to capture structural and topological relationships, whereas ELM and Naïve Bayes offer superior efficiency. SVM and LS-SVM maintain robust stability under noisy conditions, and CNNs are well-suited for sequence-based prediction tasks. By combining empirical validation, critical insights, and deployment-focused recommendations, this survey delivers decision-grade guidance that bridges theoretical understanding with real-world implementation, thus clarifying the trade-offs among accuracy, efficiency, and scalability in PPI detection research. Full article

Autoencoder-based models have become a fundamental component of unsupervised and self-supervised learning in natural language processing (NLP), enabling models to learn compact latent representations through input reconstruction. From early denoising autoencoders to probabilistic variational autoencoders (VAEs) and transformer-based masked autoencoding, reconstruction-driven objectives have played a significant role in shaping modern approaches to text representation and generation. This review provides a comprehensive analysis of the evolution of autoencoder architectures and training objectives in NLP, and synthesizes applications of VAEs across language modeling, controllable text generation, machine translation, sentiment modeling, and multilingual representation learning. Although previous surveys have examined deep generative models or representation learning in NLP, there remains a lack of a unified review that systematically connects classical autoencoder variants, variational formulations, and modern transformer-based masked autoencoders within a single conceptual framework. To address this gap, this work consolidates architectural developments, training objectives, and major application domains under a reconstruction-based learning perspective, offering a structured comparison of modeling choices, datasets, and evaluation practices. Our analysis highlights the strengths and limitations of existing approaches, discusses the ongoing influence of autoencoder-style learning in NLP, and outlines future research directions focused on improving training stability, designing more structured latent spaces, and enhancing multilingual representation learning. Full article

Structural Equation Modeling (SEM) is a key framework for analyzing complex economic relationships involving latent variables, mediation effects, and endogeneity, yet the choice between frequentist and Bayesian estimation remains theoretically and practically contested, especially in settings with non-stationary data and small samples. This study provides a formal comparison of the two approaches by formulating SEM as a probabilistic graphical model and deriving the corresponding estimation procedures, identifiability conditions, and uncertainty measures. We examine asymptotic properties of frequentist estimators and posterior consistency in Bayesian SEM, with particular attention to integrated time-series SEM applications such as shadow economy estimation. The analysis shows that while both approaches converge under large-sample conditions, important differences arise in finite samples. Bayesian methods exhibit more stable point estimates through coherent uncertainty quantification, particularly when prior information regularizes an otherwise ill-conditioned likelihood. Under model misspecification, Bayesian posteriors concentrate around the pseudo-true parameter defined by the Kullback-Leibler projection, providing a probabilistic representation of misspecification uncertainty through posterior spread—an advantage over frequentist inference, which typically conditions on the maintained model as exact. These findings carry direct implications for empirical economic modeling under realistic data constraints. In settings where sample sizes are small, identification is weak, and model uncertainty is substantial, conditions that routinely characterize macroeconomic research, the choice of inferential framework is not a matter of philosophical preference but a determinant of whether policy-relevant conclusions can be credibly defended. Bayesian SEM offers a principled and transparent path forward in precisely these conditions. Full article

This paper presents a comprehensive formulation of fuzzy path choice, based on representing utilities through fuzzy numbers. This approach advances the modelling of path choice problems in transportation systems. This model improves the ability to capture the uncertainty of travellers’ perceptions and behaviours, providing an alternative to traditional probabilistic frameworks. These models are the core of the assignment models used to simulate transport systems and calculate sustainability indicators. To support its use in assignment procedures, the paper set out the mathematical operations required for manipulating fuzzy quantities, ensuring internal consistency and operational feasibility. A key contribution is the combined use of normalised and non-normalised fuzzy numbers, which increases modelling flexibility and provides a novel way to simulate path overlap. The model is based on two approaches: the introduction of a factor that modifies the core of the fuzzy number, and an approach that modifies the confidence of the fuzzy number. The two approaches are specified and applied in a test network. Numerical applications demonstrate that the proposed method effectively accounts for path dependencies where traditional fuzzy operators fail. Full article

Geologic carbon storage (GCS) is expanding rapidly as a cornerstone decarbonization option, but its climate value depends on maintaining long-term containment of CO₂ and displaced formation brine. Legacy wells—many drilled and abandoned before modern barrier standards—remain one of the most credible and controllable pathways for unintended upward migration. To support transparent, fit-for-purpose risk screening, this study benchmarks three leakage-modeling philosophies across a common six-layer scenario: (i) a reservoir-scale analytical solution for layered aquifers, (ii) a semi-analytical pressure-transient model that captures rock–fluid compressibility and breakthrough time, and (iii) a new mechanistic wellbore-scale model that explicitly represents dominant annular failure pathways (micro-annuli, cement fractures, casing breaches, and cement–formation interface flow) with pathway-specific hydraulic losses. Results show that model choice and physics assumptions drive order-of-magnitude differences in predicted brine rates: after 1000 days, the analytical model predicts ~1.7 bbls/day, the pressure-transient model exceeds 8 bbls/day, whereas the mechanistic model yields damage-dependent outcomes (~0.2–0.4 bbls/day for moderate–severe cement damage and up to ~3.5 bbls/day for open-channel conditions). These findings demonstrate that neglecting wellbore hydraulic resistance can systematically overstate leakage risk, while mechanistic pathway representation enables more realistic, condition-dependent screening. Future work will focus on model calibration to field/monitoring data, probabilistic parameterization of defect geometries, and extension to multiphase/reactive leakage to support operational decision-making and regulatory assurance. Full article

AI-enabled chatbots are increasingly positioned as a remedy for breastfeeding support gaps in rural maternal health, offering private, immediate assistance amid persistent shortages of lactation specialists and limited access to care. However, their clinical promise remains constrained by the probabilistic nature of large language models, which can generate hallucinations that undermine maternal–infant safety. This article argues that safely integrating AI into breastfeeding support requires treating hallucination not as a singular technical flaw but as a systems-level risk shaped by design, governance, and use context. We identified key risks of AI systems that could result in hallucination such as, false citations, transcription errors, prompt injection and jailbreaking, and incorrect generalization or personalization, and analyze how each error introduces distinct safety vulnerabilities. Drawing from systems thinking, we outline mitigation strategies including retrieval-augmented generation grounded in authoritative breastfeeding sources, layered guardrails, adversarial testing, uncertainty-aware messaging, and domain-specific fine-tuning. By linking AI system design choices to downstream health consequences in resource-constrained settings, this paper reframes AI-assisted breastfeeding support as a governance challenge central to equitable, safe maternal health innovation. Full article

This study introduces a hybrid methodological approach for personalised clinical decision support, integrating SHACL-based deterministic constraints with Bayesian probabilistic models. The primary goal is to validate the model and demonstrate the benefits of combining encoded clinical knowledge with probabilistic uncertainties in managing complex therapeutic scenarios. The framework was applied to recurrent urinary tract infections (UTIs) in postmenopausal patients, a clinical context marked by high frequency, treatment challenges, and potential conflicts among therapeutic guidelines. Realistic simulated case studies were developed, encompassing both simple clinical profiles and complex situations, such as patients with antibiotic resistance. Each profile was modelled in RDF/Turtle, enabling semantic representation of clinical features and therapeutic rules. The system automatically calculates success and failure probabilities for different therapeutic scenarios, dynamically adapting them based on follow-up data. This allows clinicians to assess not only the initial therapy choice (Case study no. 1) but also the potential addition of supplementary interventions during treatment (Case study no. 2). Results highlight that the proposed hybrid SHACL–Bayesian framework enables tightly coupled deterministic–probabilistic reasoning, where SHACL constraints define the admissible clinical decisions and Bayesian inference operates within this validated space. Compared to deterministic or probabilistic approaches, the combined framework more effectively handles uncertainty, guideline conflicts, and temporal updates. The scientific contribution lies in showing that this integration enhances decision support for recurrent UTIs in postmenopausal patients, providing clinically consistent, transparent, and adaptive therapeutic recommendations aligned with the patient’s evolving condition. Full article

We present WM-Classroom v1.0, a pedagogical multi-species agent-based model (ABM) designed for educational purposes in predator–prey–harvest systems. The model embeds a predator, two prey breeds, and human harvesters on a homogeneous 50 × 50 grid with weekly time steps, implementing random movement, abstract energetics, prey consumption, reproduction, legal harvest with species-specific cut-offs and seasons, optional predator control, and a poaching switch. After basic technical checks (energetic calibration, prey composition, herbivore viability), we explore the consistency of the model under illustrative scenarios including no hunting, single-prey harvest, hunter-density and season-length gradients, predator removal, and poaching. In the no-hunting baseline (n = 100), mean end-of-run abundances were 22 deer, 159 boar, and 45 wolves, with limited extinction events. Deer-only harvest often drove deer to very low end-of-run counts (mean 1–16) with extinctions in 2–7/10 replicates across cut-offs, whereas boar-only harvest showed higher persistence (mean 11–74) and boar extinctions occurred only at the lowest cut-off (3/10). Increasing hunter numbers or season length depressed prey and could indirectly reduce wolves via prey depletion. Legal predator control reduced predators as designed, while poaching had little effect under the implemented rules. Because interaction and prey-choice rules are simplified for transparency, outcomes should be interpreted as conditional on model assumptions. WM-Classroom v1.0 provides a didactic sandbox for courses, professional training, and outreach, with extensions (habitat heterogeneity, age/sex structure, probabilistic diet/kill success, and calibration/validation) outlined for future versions. Full article

Power-system planning is being reshaped by rapid decarbonisation, electrification, and digitalisation, which collectively amplify uncertainty in demand, generation, technology adoption, and policy pathways. This review critically synthesises three principal optimisation paradigms used to plan under uncertainty in power systems: scenario-based stochastic optimisation, set-based robust optimisation (including adaptive and distributionally robust variants), and minimax-regret decision models. The review is positioned to address a recurrent limitation of many uncertainty-planning surveys, namely the separation between “method reviews” and “technology reviews”, and the consequent lack of decision-operational guidance for planners and system operators. The central contribution is a decision-centric framework that operationalises method selection through two explicit dimensions. The first is an information posture, which formalises what uncertainty information is credible and usable in practice (probabilistic, set-based, or probability-free scenario representations). The second is a flexibility posture, which formalises the availability, controllability, and timing of operational recourse enabled by smart-grid technologies. These postures are connected to modelling templates, data requirements, tractability implications, and validation/stress-testing needs. Smart-grid technologies are integrated not as an appended catalogue but as explicit sources of recourse that change the economics of uncertainty and, in turn, shift the relative attractiveness of stochastic, robust, and regret-based planning. Soft Open Points, Coordinated Voltage Control, and Vehicle-to-Grid/Vehicle-to-Building are treated uniformly under this recourse lens, highlighting how device capabilities, control timescales, and implementation constraints map into each paradigm. The paper also increases methodological transparency by describing literature-search, screening, and inclusion principles consistent with a structured narrative review. Practical guidance is provided on modelling choices, uncertainty governance, computational scalability, and institutional adoption constraints, alongside revised comparative tables that embed data credibility, regulatory interpretability, and implementation maturity. Full article

Pairs trading via futures calendar spreads offers a robust market-neutral approach to exploiting transient mispricings, yet real-time implementation is hindered by asynchronous trading. This paper introduces a Cointegrated Ising Spin Model, CISM, for real-time signal generation in high-frequency spread trading. The model links the macro-level equilibrium of cointegration with micro-level agent interactions, representing prices as magnetizations in an agent-based system. A novel $\Delta$-weighted arbitrage force dynamically adjusts agents’ corrective behavior to account for information staleness. Calibrated on tick-by-tick Brent crude oil futures, the model produces a time-varying probability of spread reversion, enabling probabilistic trading decisions. Backtesting demonstrates a 74.65% success rate, confirming the CISM’s ability to generate stable, data-driven arbitrage signals in asynchronous environments. The model bridges macro-level cointegration with micro-level agent interactions, representing prices as magnetizations within an agent-based Ising system. A novel feature is a $\Delta$-weighted arbitrage force, where the corrective pressure applied by agents in response to the standard Error Correction Term is dynamically amplified based on information staleness. The model is calibrated on historical tick data and designed to operate in real time, continuously updating its probability-based trading signals as new quotes arrive. The model is framed within the context of Discrete Choice Theory, treating agent transitions as utility-maximizing decisions within a Vector Logistic Autoregressive (VLAR) framework. Full article

It has been well established that individuals employ different decision strategies depending on the task environment, and these strategies differ in the amount of information search and time required to reach a decision. The present study developed probabilistic models for four representative decision strategies—additive, conjunctive, disjunctive, and lexicographic (including lexicographic semi-order)—and applied them to predict expected information search and decision time in multi-attribute decision-making tasks that varied in the number of attributes and alternatives. The modeling results showed that conjunctive and disjunctive strategies were strongly influenced by the number of attributes but were relatively unaffected by the number of alternatives. In contrast, the additive and lexicographic strategies were affected by both the number of attributes and alternatives, although the influence was smaller for the lexicographic strategy. To evaluate the predictive validity of these probabilistic models, their predictions were compared with those obtained through computer simulations based on an adaptive decision-maker model using the Mersenne Twister method, as well as with data from the previous psychological experiment. The comparative analyses revealed that the predictions generated by the probabilistic models were generally consistent with findings from prior empirical and simulation studies. These results suggest that even relatively simple mathematical models can successfully account for and predict variations in information search behavior and decision time leading to final choice outcomes. Full article