EQMs, sixty LLM-scored reasoning patterns, predict forecast accuracy at both item and person levels and outperform prior text-analysis methods in a large pre-registered tournament dataset.
super hub
Data Mining, Inference, and Prediction
38 Pith papers cite this work, alongside 21,949 external citations. Polarity classification is still indexing.
38
Pith papers citing it
21.9k
external citations · Crossref
hub tools
citation-role summary
background 3
method 1
citation-polarity summary
claims ledger
- background deployment because the evaluation protocol does not reflect real-world conditions or because issues such as data leakage and distribution shift are overlooked [1, 2]. This gap between apparent validation success and operational performance highlights the need for more rigorous and context-aware evaluation methods. The primary objective of model evaluation is to estimate how well a learned model generalizes to unseen data [3]. However, generalization cannot be reduced to a single universal criter
- background One way to explain these results is the basic fact that BSRBF-KAN, FastKAN, FasterKAN and MLP have more parameters than the LTBs-KAN architecture. In addition, FastKAN,and FasterKANuse a linear function as spline which allows to reduce complexity during training improving the learnability of the information in the data set. This is a phenomenon that happens in learning known as the "bias-variance dilemma" [43] which points out to the need to increase the size of the dataset when using more compl
- method For each outer foldk= 1, . . . , K out, the training data are passed to the inner loop, where cross-validation is performed to select optimal hyper- parameters ˆθk. The model trained on the corresponding inner training set is then evaluated on the outer test fold. The overall error estimate is obtained as: ˆE= 1 Kout KoutX k=1 L y(k), f(x(k); ˆθk) ,(20) whereL(·,·) denotes the loss function. In this study, the root mean squared error (RMSE) was used. As Eq. (20) indicates, the error estimate r
- background Figure 9: Hyperparameter analysis for adaptive Material Fingerprinting applied to skin data. Figure 10: Stress-stretch data and discovered model for skin with hyperparametersn a =1 ands=0.7, averageR 2 =0.3777. framework adds one anisotropic term and discovers the following transversely isotropic 2-term model ˜W= +2.3271·10 −3 3X j=2 X k<j h exp (9.08[λjλk −1]) i +7.4216·10 2 log(cosh(0.60 [λa −1])) 2. (15) This model achieves a substantially improved average accuracy ofR2 =0.8441, see Fig. 11,
authors
co-cited works
representative citing papers
Operator-adaptive PLS and Ridge models integrate linear preprocessing screening internally via algebraic identities, delivering comparable or better prediction accuracy than exhaustive external search on NIR regression and classification tasks with orders-of-magnitude lower fitting time.
Soft-MSM is a smooth, gradient-enabled version of the context-aware MSM distance for time series alignment that outperforms Soft-DTW alternatives in clustering and nearest-centroid classification.
Survival models show distinct interpolation behaviors driven by loss functions, so overparametrization does not reliably improve generalization as in regression or classification.
For generic linear concentration models the maximum likelihood threshold equals the naive dimension count.
Manifold-adapted anisotropic radial basis functions, shaped by clustering, yield a global explicit non-intrusive reduced vector field that recovers chaotic invariant measures competitively with intrusive and neural models.
Prior information in Gaussian graphical models is converted to an effective sample size, with two Bayes-factor and information-ratio sample-size planning procedures.
The saturation index S(K)=erank(pooled within-class covariance)/K tracks when few-shot labels become redundant for fixed linear probes, with within-task median Spearman ρ≈0.81 on 17 binary tasks.
Develops a unified framework representing performance metrics as smooth functionals of confusion-matrix probabilities to enable cluster-robust sandwich variance estimation for asymptotically valid confidence intervals and tests under clustered data.
Wasserstein least squares extends Euclidean least squares to distribution-valued responses via convex analysis, yielding n^{-1/2} rates under template deformation and faster barycenter rates than prior work.
Establishes n^{1-ε}-hardness of approximation for dichromatic number and acyclic number on tournaments, plus polynomial-time approximations for ℓ-dicolorable digraphs and special dense cases.
A framework using generative AI to produce synthetic multilevel data for Monte Carlo simulations that evaluate the performance and parameter recovery of quantitative methods.
LTBs-KAN delivers linear-time B-spline evaluation in KANs plus parameter reduction via product-of-sums factorization, with competitive results on MNIST, Fashion-MNIST, and CIFAR-10.
A Bayesian uncertainty-aware framework localizes tendon breakage in pre-stressed concrete using DFOS data, FEM with embedded model-form uncertainty, and Gaussian Process surrogates.
SDAMI detects interactions in high-dimensional data via an Effect Footprint principle and models them using sparsity, group lasso, and dedicated deep subnetworks for improved interpretability.
PTMC is a proposed Monte Carlo estimator that generates market-outcome distributions by simulating continuous double-auction interactions among persona-conditioned neural-policy bots whose heterogeneity is drawn from a learned distribution.
A two-step framework combines stacked hurdle random forest models for local severity prediction with semi-parametric spatio-temporal modeling to reconstruct large-scale disease dynamics from imperfect indicators, demonstrated on sugar beet yellows in France.
Joint NMF and binomial regression learns response-relevant text signals with competitive performance on simulations and review data.
Proposes PcovRnnp method enabling simultaneous dimension reduction and regularized coefficient estimation via nuclear norm penalty in high-dimensional settings.
A note that flags an oversight in RLT convergence proofs for polynomial optimization and recovers correctness via one extra natural assumption.
Introduces MADQI, a label-free composite metric combining ARC, PPS, SDS, and ECE to evaluate unsupervised anomaly detection in AIS vessel data, reporting 80.37% on tested datasets.
DCP integrates distribution-generating predictors with score-agnostic conformal calibration via numerical inversion to produce valid and efficient time-series prediction intervals, supported by a modified Winkler score that penalizes undercoverage.
Machine learning framework for emerald grading reaches 98% accuracy on a new public dataset of 192 images and outperforms a deep learning baseline.
The paper factorizes HNDL compromise probability into temporal hazard, multiplicative vulnerability-exposure terms, and a saturation factor justified by assumptions on adversarial production and value decay.
citing papers explorer
-
Measuring Judgment Quality in Natural-Language Explanations: Evidence from Forecasting Tournaments
EQMs, sixty LLM-scored reasoning patterns, predict forecast accuracy at both item and person levels and outperform prior text-analysis methods in a large pre-registered tournament dataset.
-
Reframing preprocessing selection as model-internal calibration in near-infrared spectroscopy: A large-scale benchmark of operator-adaptive PLS and Ridge models
Operator-adaptive PLS and Ridge models integrate linear preprocessing screening internally via algebraic identities, delivering comparable or better prediction accuracy than exhaustive external search on NIR regression and classification tasks with orders-of-magnitude lower fitting time.
-
Soft-MSM: Differentiable Context-Aware Elastic Alignment for Time Series
Soft-MSM is a smooth, gradient-enabled version of the context-aware MSM distance for time series alignment that outperforms Soft-DTW alternatives in clustering and nearest-centroid classification.
-
Understanding Overparametrization in Survival Models through Interpolation
Survival models show distinct interpolation behaviors driven by loss functions, so overparametrization does not reliably improve generalization as in regression or classification.
-
Maximum likelihood thresholds of generic linear concentration models
For generic linear concentration models the maximum likelihood threshold equals the naive dimension count.
-
Manifold-adapted radial basis functions for reduced-order modelling of chaotic flows
Manifold-adapted anisotropic radial basis functions, shaped by clustering, yield a global explicit non-intrusive reduced vector field that recovers chaotic invariant measures competitively with intrusive and neural models.
-
What is your Prior Worth? Effective Sample Size and Sample Size Planning for Gaussian Graphical Models
Prior information in Gaussian graphical models is converted to an effective sample size, with two Bayes-factor and information-ratio sample-size planning procedures.
-
The Geometry of Saturation: Effective Rank Predicts When Labels Stop Helping in Few-Shot Classification
The saturation index S(K)=erank(pooled within-class covariance)/K tracks when few-shot labels become redundant for fixed linear probes, with within-task median Spearman ρ≈0.81 on 17 binary tasks.
-
Beyond Point Estimates: Reliable Evaluation of Prediction Performance Metrics under Clustered Data
Develops a unified framework representing performance metrics as smooth functionals of confusion-matrix probabilities to enable cluster-robust sandwich variance estimation for asymptotically valid confidence intervals and tests under clustered data.
-
Wasserstein Least Squares: A Canonical Regression Method for Probability Distributions
Wasserstein least squares extends Euclidean least squares to distribution-valued responses via convex analysis, yielding n^{-1/2} rates under template deformation and faster barycenter rates than prior work.
-
Hardness and Approximation for Coloring Digraphs
Establishes n^{1-ε}-hardness of approximation for dichromatic number and acyclic number on tournaments, plus polynomial-time approximations for ℓ-dicolorable digraphs and special dense cases.
-
Generative AI-Based Monte Carlo Simulation for Method Evaluation Using Synthetic Multilevel Data
A framework using generative AI to produce synthetic multilevel data for Monte Carlo simulations that evaluate the performance and parameter recovery of quantitative methods.
-
LTBs-KAN: Linear-Time B-splines Kolmogorov-Arnold Networks
LTBs-KAN delivers linear-time B-spline evaluation in KANs plus parameter reduction via product-of-sums factorization, with competitive results on MNIST, Fashion-MNIST, and CIFAR-10.
-
Bayesian Tendon Breakage Localization under Model Uncertainty Using Distributed Fiber Optic Sensors
A Bayesian uncertainty-aware framework localizes tendon breakage in pre-stressed concrete using DFOS data, FEM with embedded model-form uncertainty, and Gaussian Process surrogates.
-
Sparse Deep Additive Model with Interactions: Enhancing Interpretability and Predictability
SDAMI detects interactions in high-dimensional data via an Effect Footprint principle and models them using sparsity, group lasso, and dedicated deep subnetworks for improved interpretability.
-
Persona-Trained Monte Carlo: Estimating Market-Outcome Distributions via Swarms of Persona-Conditioned Neural Policy Bots in a Limit Order Book
PTMC is a proposed Monte Carlo estimator that generates market-outcome distributions by simulating continuous double-auction interactions among persona-conditioned neural-policy bots whose heterogeneity is drawn from a learned distribution.
-
Predicting disease severity and large-scale spread from coupled severity measurements and imperfect indicators: Application to beet yellows
A two-step framework combines stacked hurdle random forest models for local severity prediction with semi-parametric spatio-temporal modeling to reconstruct large-scale disease dynamics from imperfect indicators, demonstrated on sugar beet yellows in France.
-
Learning Interpretable Text Signals for Structured Responses
Joint NMF and binomial regression learns response-relevant text signals with competitive performance on simulations and review data.
-
Principal Covariate Regression with Nuclear Norm Penalty
Proposes PcovRnnp method enabling simultaneous dimension reduction and regularized coefficient estimation via nuclear norm penalty in high-dimensional settings.
-
A note on the convergence guarantees of RLT-based algorithms for polynomial optimization
A note that flags an oversight in RLT convergence proofs for polynomial optimization and recovers correctness via one extra natural assumption.
-
A Novel Evaluation Metric for Unsupervised Learning in AIS-Based Maritime Anomaly Detection: MADQI
Introduces MADQI, a label-free composite metric combining ARC, PPS, SDS, and ECE to evaluate unsupervised anomaly detection in AIS vessel data, reporting 80.37% on tested datasets.
-
Distribution-Aware Conformal Prediction: A Framework for generating efficient prediction intervals for time series
DCP integrates distribution-generating predictors with score-agnostic conformal calibration via numerical inversion to produce valid and efficient time-series prediction intervals, supported by a modified Winkler score that penalizes undercoverage.
-
Machine learning applied to emerald gemstone grading: framework proposal and creation of a public dataset
Machine learning framework for emerald grading reaches 98% accuracy on a new public dataset of 192 images and outperforms a deep learning baseline.
-
A Formal Basis for Quantum Cryptographic Exposure Measurement under HNDL Threat
The paper factorizes HNDL compromise probability into temporal hazard, multiplicative vulnerability-exposure terms, and a saturation factor justified by assumptions on adversarial production and value decay.
-
Efficient event-driven retrieval in high-capacity kernel Hopfield networks
Asynchronous sequential updates in KLR Hopfield networks produce statistically indistinguishable trajectories from synchronous dynamics, achieve empirical capacities near P/N=30, and converge with event counts close to initial Hamming distance.
-
Audio Video Verbal Analysis (AVVA) for Capturing Classroom Dialogues
AVVA is a new framework adapting verbal analysis for classroom discourse with triangulation across ten steps and a four-criterion validation scheme for temporal stability, applied to 23 hours of recordings.
-
Resonance Statistics -Informed Fitting Applied to Automated Cross Section Evaluation
Resonance statistics-informed methods in automated fitting reduce spin group bias, enhance Wigner statistics consistency, and stabilize resonance density with minimal impact on cross section fit quality.
-
Adaptive Material Fingerprinting for the fast discovery of polyconvex feature combinations in isotropic and anisotropic hyperelasticity
Adaptive Material Fingerprinting builds hyperelastic material models as greedy linear combinations of precomputed feature fingerprints, reaching neural-network-level fit accuracy on rubber and skin without online optimization.
-
When Do Data-Driven Systems Exhibit the Capability to Infer?
A new framework grades levels of inference capability in data-driven systems to assess compliance with the EU AI Act definition of AI, illustrated via credit scoring workflows.
-
Assessing model calibration with boosting trees
Boosting trees test necessary conditions for calibration and auto-calibration of regression models, shown powerful on a large insurance dataset.
-
Revealing Geography-Driven Signals in Zone-Level Claim Frequency Models: An Empirical Study using Environmental and Visual Predictors
Augmenting zone-level MTPL claim frequency models with coordinates, environmental features at 5 km scale, and image embeddings improves predictive accuracy on unseen postcodes across GLM, regularized GLM, and tree-based models.
-
ClinQueryAgent: A Conversational Agent for Population Health Management
The paper introduces ClinQueryAgent, a conversational agent that converts natural language queries into database queries for population health management while keeping patient data secure, and reports its use by 128 staff across 15 NHS practices covering 148,319 patients.
-
Autoencoder Architectures for Athlete Performance Scoring from Wearable Telemetry
Deep autoencoders outperform PCA and VAE variants on a composite of reconstruction MSE and interpretability metrics when reducing runner wearable data to a single latent performance score.
-
Smart Ensemble Learning Framework for Predicting Groundwater Heavy Metal Pollution
Ensemble learning with Gaussian copula transformation predicts groundwater heavy metal pollution index with high accuracy (R²=0.96) while identifying key contaminants via clustering.
-
Evaluating Supervised Machine Learning Models: Principles, Pitfalls, and Metric Selection
Model evaluation in supervised learning should be treated as a context-dependent, decision-oriented process aligned with operational objectives rather than relying on a small set of aggregate metrics.
-
Bias in Large Language Models: Origin, Evaluation, and Mitigation
A literature review that categorizes bias in LLMs, surveys evaluation and mitigation techniques, and discusses ethical implications.
- The Good, the Bad, and the Ugly of Markov Boundary for Tabular Prediction
- Computing k-means in mixed precision