Renewable-generation forecasting is a probabilistic time-series problem in which point accuracy alone is insufficient for operational decision-making. This study proposes a physics-informed Bayesian framework for multi-horizon forecasting of solar, wind, and total renewable generation. The task is formulated as a 15-dimensional target–horizon problem covering three generation families and five forecast horizons: H1, H3, H6, H12, and H24. A leakage-safe data construction protocol generates causal predictors from meteorological observations, generation history, calendar cycles, lagged and rolling statistics, ramp descriptors, and physics-informed transformations. PI-BHTF partitions the 422-dimensional predictor space into solar, wind, temporal-calendar, and cross-context components, encodes them through parallel nonlinear branches, and combines the representations using cross-energy gated fusion. Its neural core uses a heteroscedastic predictive head and Monte Carlo dropout to distinguish input-dependent aleatoric uncertainty from epistemic variability, whereas the final hybrid forecasts are calibrated using residual quantiles computed from a chronologically held-out validation segment. Across three prespecified random seeds, the full PI-BHTF achieved a mean MAE of 0.0797 ± 0.0007 on the internal chronological test and 0.0686 ± 0.0006 on locked external SCADA validation. Performance and calibration varied substantially across target–horizon tasks, including marked short-horizon external undercoverage for wind generation. Component-wise ablation supported semantic feature partitioning and the heteroscedastic head, whereas the physics-informed features, cross-energy gate, and physics-consistency loss did not independently reduce aggregate MAE. PI-BHTF should therefore be interpreted as a reproducibly competitive framework that balances multi-horizon accuracy, structured representation, uncertainty estimation, and external transferability rather than as a universally dominant model.

The article presents an annotated corpus consisting of clinical texts in Russian, obtained by exomic sequencing. This corpus was developed to support the tasks of automatically identifying named objects and semantic relationships in relation to genes, mutations, hereditary diseases, phenotypic traits and their clinical significance. During the formation of the corpus, reports of actual clinical exomic sequencing were used, the data went through the stages of preliminary anonymization and text normalization. The labeling process used international standards and knowledge bases such as HGVS, OMIM, ClinVar, and HPO, and ensured consistency and accuracy of biomedical information. The corpus contains more than 25,000 biomedical objects and more than 6,000 semantic links, making it an important resource in the field of clinical genetics in terms of volume and content. The annotation was carried out manually with the participation of several experts, and the results were compared by cross-checking, and the level of agreement between the annotators was assessed using special indicators. The results obtained indicate the high quality and reliability of the case. The finished corpus makes it possible to effectively use natural language processing models in the field of medical genetics for teaching and evaluation, development of clinical decision support systems, and applied research for structuring genetic data.

Allergic sensitization is commonly modeled as a binary classification outcome, an approach that discards the probabilistic structure linking demographic, immunological, and clinical variables and offers limited support for individualized risk communication. This work develops and evaluates a Bayesian network (BN) for probabilistic stratification of laboratory-defined serological IgE sensitization, fitted on harmonized data from the National Health and Nutrition Examination Survey (NHANES) 2005–2006 cycle (analytic cohort n = 8,285). The network integrates demographic covariates (age, sex), immunological markers (total serum IgE, eosinophil percentage), and clinical indicators (physician-diagnosed asthma, self-reported allergic symptoms) into a directed acyclic graph whose conditional probability tables were estimated by maximum-likelihood estimation with Laplace smoothing. Serological IgE sensitization, defined as at least one allergen-specific IgE ≥ 0.35 kUA/L across a 19-allergen panel, was present in 47.1% of the cohort. A five-fold cross-validated comparison of BN, logistic regression (LR), and random forest (RF) on identical predictor sets showed indistinguishable discrimination: AUC 0.808/0.809/0.808 (IgE + eosinophils) and 0.820/0.824/0.822 (IgE + eosinophils + age + sex). Individual-level calibration was excellent: Brier score = 0.171, calibration slope = 0.9998 (95% CI 0.9563–1.0434), intercept = −0.0001 (95% CI − 0.0522 to 0.0520). The three posterior probability strata (low < 0.40, intermediate 0.40–0.70, high > 0.70) correspond structurally to the three total IgE tertiles; adding eosinophil category improves continuous discrimination by ΔAUC = +0.010 but reclassifies zero participants between strata (categorical NRI = 0). Bidirectional inference—updating sensitization probability from a confirmed asthma diagnosis—raised the posterior from 0.472 to 0.651 (ΔAUC +0.006, asthma subsample n = 8,275), confirming a structural advantage unavailable from unidirectional classifiers. Observed serological sensitization rates across strata were 13.8% (low), 46.1% (intermediate), and 81.6% (high). The framework is offered as a transparent, auditable methodological demonstration of probabilistic modeling of serological IgE sensitization. It does not model clinical allergy and should not be used to guide clinical decisions without prospective validation. Its principal advantages over discriminative classifiers are structural: bidirectional inference, modular auditability of conditional probability tables, and native uncertainty propagation.

The article examines the influence of Markov processes on computations in Bayesian networks (BN), an important area of research within probabilistic graphical models. The concept of Bayesian Markov networks (BMN) is introduced, an extension of traditional Bayesian networks with the addition of a Markov constraint, according to which the probability in a node can only depend on the state of neighboring nodes. This constraint makes the model more realistic for many practical tasks, as most graphical models that reflect real-world processes possess the Markov property. The article also discusses that Bayesian networks, in the absence of evidence, actually exhibit the Markov property. However, when evidence (additional information) is introduced into the model, challenges arise that require more complex computational methods. In response, the article proposes algorithms adapted for working with Bayesian Markov networks in the presence of evidence. These algorithms are aimed at optimizing computations and reducing computational complexity. Additionally, a comparative analysis of calculations in Bayesian networks without Markov constraints and with them is conducted, highlighting the advantages and disadvantages of each approach. Special attention is paid to the practical applications of the proposed methods and their effectiveness in various scenarios.

Integrating newly emerging terms into existing ontologies is a recurring maintenance problem in knowledge engineering, particularly in biomedical domains where terminology evolves faster than manual curation can accommodate. This paper addresses the placement of concepts that are absent from the target ontology—the out-of-knowledge-base setting—in which a textual mention must be assigned one or more insertion positions in the subsumption hierarchy rather than linked to an existing node. We propose a three-stage framework that extends the conventional retrieve-then-select paradigm with an intermediate stage of edge generation and enrichment, which expands the candidate set by traversing the local structure of the ontology. Stage 1 retrieves candidate edges using a fine-tuned bi-encoder trained with a max-margin objective; Stage 2 constructs and structurally enriches candidate edges; Stage 3 selects among them using either a fine-tuned cross-encoder or a large language model under explainable instruction tuning. We evaluate on two datasets derived from SNOMED CT, MM-S14-Disease and MM-S14-CPP, under a strict out-of-knowledge-base protocol. Fine-tuned pre-trained language models outperform zero-shot and instruction-tuned large language models on ranking accuracy, while the instruction-tuned configuration produces expert-auditable justifications at a modest cost in accuracy. On MM-S14-Disease, the strongest configuration places a correct insertion edge among the ten highest-ranked candidates for 38.7% of test mentions and recovers the complete gold edge set for 16.4%, against 26.1% and 9.2% for retrieval alone. The framework is positioned as decision support for ontology curators rather than as an autonomous ontology generator.

The growing volume of unstructured clinical and genetic text calls for automated processing methods that can meaningfully support clinical decision-making. Most existing work, however, treats the analysis of patient clinical descriptions and the interpretation of genetic reports as unrelated tasks, disregarding the sequence in which real diagnostic decisions unfold. We study two stage-specific families of classification tasks corresponding to the two points at which such decisions are taken—pre-genetic triage from clinical narratives, and post-genetic interpretation of completed genetic reports—and evaluate each family under a single leakage-controlled protocol using TF-IDF text representation with classical machine learning. All experiments use a corpus of 546 records with confirmed provenance, partitioned by a group-constrained master split (349 train/88 validation/109 test, seed 42) in which all 181 text-similarity candidate pairs are confined to a single partition. Task subsets contain 299 and 326 documents at the first stage and 375 and 461 at the second. Text representation relied on Word TF-IDF, Character TF-IDF and combined Word + Character TF-IDF with Logistic Regression and Linear SVM; every configuration was selected on validation Macro F1 alone, with the test partition opened only after selection. RuBERT-tiny2, XLM-RoBERTa base and PubMedBERT were fine-tuned as transformer baselines over three seeds each. Under this protocol, the first stage reached Macro F1 = 0.6876 (Accuracy = 0.8308, ROC-AUC = 0.8500) and the second-stage Macro F1 = 0.8225 (Accuracy = 0.8571, ROC-AUC = 0.9377) for the binary classification of diagnostic status. A controlled comparison holding the documents, split, vectorizer, classifier and hyperparameters fixed and varying only the label version increased Accuracy from 0.5738 to 0.8361 (exact McNemar p = 0.000145), whereas Macro F1 increased numerically from 0.5165 to 0.5966 but not significantly (paired bootstrap p = 0.354; 95% CI of the difference [−0.091, 0.229]). Two Russian and multilingual encoders collapsed to the majority class (Macro F1 = 0.4583), while an English biomedical encoder scored higher than the classical comparator (0.7705 ± 0.0119 against 0.6876) without reaching significance (paired bootstrap p = 0.463). Thus, the controlled experiment supports improved overall correctness, mainly associated with the majority class, but does not establish improved balanced class-wise performance at this sample size. In the second-stage leakage analysis, removing the diagnostic conclusion reduced Macro F1 from 0.8225 to 0.6478 (paired bootstrap p = 0.001), masking the exact label-generating rules reduced it to 0.7455 (p = 0.021), and combined masking reduced it to 0.6389 (p < 0.001). The unmasked Stage 2 score therefore depends materially on explicit report cues and should not be interpreted as evidence of independent diagnostic reasoning. The contribution of this work is a stage-oriented formulation of clinical genetic text classification in which the target variable of each task is aligned with the information available at the corresponding point of the diagnostic pathway, evaluated under a reproducible leakage-controlled protocol with bootstrap confidence intervals and paired significance testing. A unified single-model baseline on the joint target reached Macro F1 = 0.5238, below either stage-specific model on its own target. Because the labels are rule-derived rather than independently expert-validated, the corpus is small, and no external validation was performed, these results characterize what is achievable on this corpus rather than demonstrating clinical readiness.