Biodata Mining最新文献

{"title":"A biology-based quality-diversity algorithm for drug repurposing in Alzheimer's disease using automated machine learning.","authors":"Sisi Shao, Pedro Henrique Ribeiro, Alena Orlenko, Katie M Cardone, Christina M Ramirez, Li Shen, Marylyn D Ritchie, Jason H Moore","doi":"10.1186/s13040-026-00550-4","DOIUrl":"10.1186/s13040-026-00550-4","url":null,"abstract":"<p><p>BACKGROUND: Alzheimer’s disease (AD) remains a major therapeutic challenge, characterized by high clinical trial failure rates and limited efficacy of current treatments. Drug repurposing offers a faster, lower-risk route to new therapies; however, existing computational approaches often prioritize predictive accuracy over mechanistic novelty and interpretability, both of which are critical for clinical translation. RESULTS: We introduce a quality-diversity Automated Machine Learning (AutoML) framework that integrates biologically informed graph neural network (GNN) embeddings with a MAP-Elites-guided search to discover predictive yet mechanistically distinct therapeutic hypotheses. Drugs and genes are embedded using GraphSAGE and variational graph autoencoders trained on the Alzheimer’s Knowledge Base (AlzKB), with a clustering loss used to anchor known AD entities and define dimensions of biological novelty. In an AD case study using matched ADSP GWAS-derived features, our framework successfully recovered known drug–gene relationships and identified robust consensus candidates across independent validation runs. Most notably, the search consistently prioritized Triclosan—a recently identified environmental risk factor for AD neuroinflammation—and the Ketamine/Quazepam pair, suggesting a model-driven preference for restoring synaptic E/I balance. Furthermore, exploratory leads such as Exemestane and Felodipine were identified in underrepresented biological niches, supported by enrichment in oxidative stress and autophagy pathways. The framework demonstrated high stability across multiple random seeds and a 48% reduction in computational cost compared to standard multi-objective evolutionary baselines. CONCLUSIONS: Beyond AD, this framework offers a generalizable strategy for integrating biomedical knowledge graphs with diversity-enhancing AutoML to accelerate the discovery of mechanistically novel drug candidates across complex polygenic diseases. </p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13159293/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147582189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Harnessing machine learning with auditory tests and demographic factors to forecast children's reading abilities in children living with and without HIV.","authors":"Jeff Joseph, Christopher Niemczak, Jonathan Lichtenstein, Albert Magohe, Samantha Leigh, Linda Zhang, Enica Massawe, Jiang Gui, Jay C Buckey","doi":"10.1186/s13040-026-00545-1","DOIUrl":"10.1186/s13040-026-00545-1","url":null,"abstract":"<p><p>Early identification of children at risk for reading difficulties is crucial to address HIV-related educational deficits and prevent long-term challenges, especially in populations affected by factors that impair cognitive development. Central auditory tests (CATs), which assess how the brain processes complex sounds, are strongly linked to reading ability and may help predict literacy outcomes. This study used four preprocessing methods to adjust for age-related variation in CAT scores, then applied machine learning to predict oral reading fluency in a longitudinal cohort of 251 children living with HIV (mean age 4.3 years − 50% female) and without HIV (mean age 4.8 years − 55% female). Early CAT performance combined with demographic factors (education quality, gender, HIV status) predicted later reading ability effectively. While predictions were stronger in the full sample than in HIV subgroups, this approach shows promise and lays the foundation for early literacy screening.</p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13151343/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147576157","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DBSCAN applied to EHRs data from patients with glioblastoma clusters patients based on cytosolic Hsp70 protein, sex, and brain subventricular zone.","authors":"Davide Chicco, Srinjoy Dora, Luca Oneto","doi":"10.1186/s13040-026-00549-x","DOIUrl":"10.1186/s13040-026-00549-x","url":null,"abstract":"<p><p>Glioblastoma is an aggressive brain cancer that kills approximately one hundred thousand people worldwide every year. Unfortunately, treatment and therapy for patients with this disease are complicated and have limited efficacy in improving individuals’ chances of survival. Electronic health records (EHRs) contain patient information collected routinely at hospitals through medical visits and laboratory tests, providing an interesting source of data for computational analyses. Clustering is an area of unsupervised machine learning where an algorithm partitions data according to certain statistical properties or rules, thereby identifying hidden patterns and correlations that would otherwise be difficult to notice. In this study, we applied several clustering techniques to three open datasets (Munich2019, Tainan2020, and Utrecht2019) derived from electronic health records, which included clinical, genetic, and administrative features of patients diagnosed with glioblastoma, considering two possible clusters. We evaluated our clustering results with the Density-Based Clustering Validation (DBCV) index, a relatively new score capable of accurately assessing both convex-shaped and concave-shaped clusters. Among the methods tested, Density-based Spatial Clustering of Applications with Noise (DBSCAN) yielded the best results across all three datasets. We then analyzed the features of the clusters identified by DBSCAN and found that cytosolic Hsp70 protein in the Munich2019 dataset, sex in the Tainan2020 dataset, and brain subventricular zone in the Utrecht2019 resulted significantly capable to distinguish the two clusters.</p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13147812/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147522449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Clustering-based low-rank matrix approximation for multimodal medical image compression.","authors":"Sisipho Hamlomo, Marcellin Atemkeng","doi":"10.1186/s13040-026-00541-5","DOIUrl":"10.1186/s13040-026-00541-5","url":null,"abstract":"<p><p>Medical images are inherently high-resolution and contain locally varying structures that are crucial for diagnosis. Efficient compression of such data must therefore preserve diagnostic fidelity while minimizing redundancy. Low-rank matrix approximation (LoRMA) techniques have shown strong potential for image compression by capturing global correlations; however, they often fail to adapt to local structural variations across regions of interest, i.e., tumor areas or other diagnostically significant regions. To address these limitations, we introduce an adaptive LoRMA, which partitions a medical image into overlapping patches, groups structurally similar patches into several clusters using k-means, and performs SVD within each cluster. We derive the overall compression factor accounting for patch overlap and analyze how patch size influences compression efficiency and computational cost. While the proposed adaptive LoRMA method is applicable to any data exhibiting high local variation, we focus on medical imaging due to its pronounced local variability. We evaluate and compare our adaptive LoRMA against global SVD across four imaging modalities: MRI, ultrasound, CT scan, and chest X-ray. An ablation study against block-based SVD is further conducted to isolate the contribution of similarity-based clustering, and a contextual comparison with JPEG2000 is included to situate the proposed approach within established medical image compression frameworks. Results demonstrate that adaptive LoRMA effectively preserves structural integrity, edge details, and diagnostic relevance, as measured by average peak signal-to-noise ratio (PSNR), average structural similarity index (SSIM), average mean squared error (MSE), average intersection over union (IoU), and edge preservation index (EPI). Adaptive LoRMA significantly minimizes block artifacts and residual errors, particularly in pathological regions, consistently outperforming global SVD in terms of PSNR, SSIM, IoU, EPI, and achieving lower MSE. Adaptive LoRMA prioritizes clinically salient regions while allowing aggressive compression in non-critical regions, optimizing storage efficiency. Although adaptive LoRMA requires higher processing time, its diagnostic fidelity justifies the overhead for high-compression applications.</p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13134349/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147516154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-output LSTM-based prediction of postoperative delirium: integrating baseline and perioperative data for enhanced risk stratification in older spine surgery patients.","authors":"Jungmin You, Jeongmin Kim, Jeongeun Choi, Bon-Nyeo Koo, Hyangkyu Lee","doi":"10.1186/s13040-026-00531-7","DOIUrl":"10.1186/s13040-026-00531-7","url":null,"abstract":"<p><strong>Introduction: </strong>Postoperative delirium (POD) adversely affects clinical outcomes among older adults undergoing spine surgery. However, existing predictive models often neglect multidimensional nature of delirium, including its clinical subtype, duration, severity, and timing. This study developed a multi-output Long Short-Term Memory (LSTM) neural network that integrates preoperative baseline characteristics and intraoperative acute stressors to predict multiple clinical dimensions of POD in elderly patients undergoing spinal surgery.</p><p><strong>Methods: </strong>This prospective observational study included 536 patients aged 70 or older who underwent elective spine surgery between November 2019 and May 2023. Comprehensive assessments were conducted during both the preoperative and intraoperative phases. The multi-output LSTM model incorporated preoperative baseline variables (demographic, frailty scores, cognitive function, medication count, and laboratory parameters) and intraoperative data (surgical invasiveness, duration of surgery and anesthesia, intraoperative fluid management, immediate postoperative medication use). Outcomes comprised delirium occurrence, subtype, duration, severity, and onset timing. Model performance was evaluated via accuracy, precision, recall, F1-score, and ROC curve analyses. SHapley Additive exPlanations (SHAP) analysis enhanced clinical interpretability.</p><p><strong>Results: </strong>Using solely preoperative baseline data, the model demonstrated strong predictive performance with an overall AUC of 0.76, particularly for delirium occurrence (AUC = 0.68), the duration (AUC = 0.80), and severity (AUC = 0.79). Incorporating intraoperative data substantially enhanced model performance, increasing the overall AUC to 0.81, notably improving predictions for delirium subtype (AUC up to 0.84), duration (AUC = 0.81), and onset timing (AUC up to 0.87). SHAP analysis consistently identified frailty, polypharmacy, cognitive impairment, nutritional deficiencies, and acute perioperative factors-such as surgical invasiveness, pain management-as pivotal predictors across delirium dimensions.</p><p><strong>Conclusion: </strong>The proposed multi-output LSTM model predicted multiple clinical dimensions of postoperative delirium, highlighting baseline health status as a primary determinant. Strategic integration of comprehensive baseline assessments with acute perioperative data substantially enhances predictive accuracy, informing personalized delirium prevention and management strategies for improved perioperative outcomes in older spine surgery patients.</p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13137560/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147516116","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Profile-guided Hybrid Approach for block-wise missing data handling in multi-omics: a breast cancer case study.","authors":"Esraa Hamdi Abdelaziz, Eman Amin, Rasha Ismail, Mai Mabrouk","doi":"10.1186/s13040-026-00530-8","DOIUrl":"10.1186/s13040-026-00530-8","url":null,"abstract":"<p><p>BACKGROUND: Block-wise missingness is a common challenge in multi-omics data, hindering the development of robust and generalizable machine learning models, as real-world cohorts rarely contain complete omic profiles. Many current methods either discard incomplete samples, use available-case models that need retraining when faced with new missingness patterns, or depend on full-dataset imputation, which can risk biological integrity and model stability. METHODS: Using a complete four-omics breast cancer dataset (705 patients, 1,937 features), up to 60% block-wise missingness was simulated across five clinically relevant scenarios and used to compare four strategies for handling missing data: an Imputation-Based model, Dynamic and Exhaustive Available-Case approaches, and the proposed Hybrid Approach that combines profile-guided modeling with selective, test-time imputation. Performance was evaluated using accuracy, F1 score, balanced accuracy, inference time, and variability across 15 random seeds, with significance assessed using the Wilcoxon signed-rank test. RESULTS: The Hybrid Approach consistently achieved the strongest and most stable performance. Relative to the complete-data baseline, it reached an average accuracy of 103.7%, F1 score of 123.3%, and balanced accuracy of 104.8%, outperforming the Imputation-Based method and matching or exceeding both Dynamic and Exhaustive Available-Case strategies. Statistical testing confirmed that these improvements were significant. The method also demonstrated fast and predictable inference (~ 2 s) and an average total runtime of ~ 49 s per configuration—nearly three times faster than the Exhaustive approach (~ 124 s)—while maintaining high reproducibility and low variance across seeds, a key indicator of computational stability. CONCLUSION: By selectively combining lightweight imputation with profile-specific modeling, the Hybrid Approach provides a computationally efficient and statistically robust solution for block-wise missing data. This framework offers a generalizable strategy for multi-omics data mining, and lays the foundation for future systems incorporating cross-profile learning and advanced imputation. </p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13063710/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147491963","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TLEUDS: a cascade Dual-Transfer learning system with quality- and knowledge-enhanced for precise fetal CHD screening.","authors":"Yuxuan Jiang, Jiajie Tang, Fanfan Zhu, Yuzhou Zeng, Junbo Wu, Wanqi Wang, Yuan Liu, Hongying Wang, Long Lu","doi":"10.1186/s13040-026-00535-3","DOIUrl":"10.1186/s13040-026-00535-3","url":null,"abstract":"<p><p>Congenital heart disease (CHD) is the most common major congenital anomaly and a leading cause of perinatal mortality and long-term neurodevelopmental disability globally, among which severe ductal dependent CHD (D-CHD) requires urgent clinical intervention due to the high risk of catastrophic cardiovascular collapse if undiagnosed. Despite the advancement of fetal echocardiography-based analysis models, there remains a lack of comprehensive and accurate evaluation approaches for D-CHD, especially for the precise differentiation of critical subtypes including coarctation of the aorta (CoA), interrupted aortic arch (IAA), and transposition of the great arteries (TGA). To address this gap, we developed the D-CHD Precision Screening System (TLEUDS) using 9,142 fetal aortic arch echocardiographic long-axis view images and 58 ultrasound clips from three medical institutions. Aiming to solve the challenges of uneven quality consistency and long-tail distribution in real-world ultrasound imaging data, TLEUDS adopts a cascaded Dual-Transfer learning framework for quality and knowledge enhancement, which implements a closed-loop screening process from initial disease detection (via the TEN-S module, a transfer version of EfficientNetV2-Screening) to detailed subtype-specific malformation screening (via the TY-PS module, a transfer version of YOLOv11-Precise Screening). Experimental results show that TLEUDS achieved a sensitivity of 0.988, 0.941, 0.966, and 0.900 for D-CHD overall, CoA, IAA, and TGA respectively, with corresponding specificities of 0.988, 0.987, 1.000, and 0.956, demonstrating state-of-the-art (SOTA) performance compared with general models. TLEUDS is expected to serve as a potential automated screening tool for tiered healthcare and computer-aided diagnosis of fetal D-CHD. Furthermore, the proposed quality and knowledge-enhanced cascaded Dual-Transfer learning framework fully leverages the “image quality distribution” and “structural prior knowledge” of fetal echocardiography, holding great potential for extension to other ultrasound image analysis domains.</p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13235188/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147482142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"BPX-Net: biomarker-preserved explainable networks for disease diagnosis and prognosis.","authors":"Jun Wang, Songchang Chen, Ru Wen, Haochao Ying, Wenqiu Xu, Lin Yin, Xiaojuan Deng, Can Han, Qun Zhu, Bin Zhang, Hongyan Tong, Chen Liu, Wei Chen, Jie Jin, Kai Jin, Chenming Xu, Hefeng Huang, Huafeng Wang, Dahong Qian","doi":"10.1186/s13040-026-00537-1","DOIUrl":"10.1186/s13040-026-00537-1","url":null,"abstract":"<p><p>Precise diagnosis of complex diseases increasingly depends on the integration of multimodal data. However, the high dimensionality of such data makes it difficult for traditional modeling methods to efficiently identify biomarkers, while deep learning approaches, despite their strong predictive power, often lack interpretability. Here, we introduce BPX-Net, a generalizable deep learning framework that learns compact multimodal representations through a biomarker-preserving dropout mechanism, where dropout probabilities are modulated by feature importance scores self-learned from a built-in interpretability module and, when available, further informed by clinical priors. This design enables BPX-Net to make robust predictions during inference by selectively attending to a sparse set of disease-informative predictors, making it insensitive to missing values in less relevant or redundant variables. Across multi-center cohorts covering diverse clinical tasks (disease diagnosis, prognosis, treatment response prediction, and risk stratification), BPX-Net yields substantial performance gains, e.g. achieving an average AUC of 85.43% across four tasks, outperforming baselines by 4% to 20% in the presence of missing data. More importantly, it identifies predictors aligned with established clinical knowledge. Cross-hospital validation further confirms the robustness and clinical relevance of these predictors. Collectively, BPX-Net offers a clinically grounded deep learning framework for multimodal analysis, intrinsically robust to data incompleteness and equipped with built-in interpretability, thereby eliminating reliance on computationally intensive post-hoc tools such as SHAP.</p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13097947/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147445595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multi-metric evaluation of readability in psychiatric discharge summaries.","authors":"Fuchen Li, Casey Overby Taylor, Ayah Zirikly","doi":"10.1186/s13040-026-00528-2","DOIUrl":"10.1186/s13040-026-00528-2","url":null,"abstract":"<p><p>BACKGROUND: Readability in medical notes is critical for patient engagement and comprehension, especially in mental health, where readable documentation reduces stigmatizing language, promotes patient-friendly communication, and strengthens trust. The importance of readability has been amplified by the federal OpenNotes policy (mandated by the 21st Century Cures Act since April 2021), which requires providers to offer patients immediate electronic access to their clinical notes. This mandate makes it imperative that mental health documentation is both clinically useful and easily understood. METHODS: This study evaluated psychiatric discharge summaries from the MIMIC-IV dataset. Specifically, we compared patient-facing “Discharge Instructions” against provider-facing “Brief Hospital Course” sections to assess documentation practices. We analyzed 1,745 notes associated with four major psychiatric diagnoses (Major Depression, Bipolar Disorder, Schizophrenia, and Eating Disorders), with a paired subset of 880 notes containing both sections. Readability was assessed using a multidimensional framework of 12 metrics, spanning traditional grade-level formulas (e.g., SMOG, KFGL) and deep learning–based models (ClinicalBERT, MedReadMe). The internal correlation across these metrics was also evaluated. RESULTS: Discharge Instructions were significantly shorter, structurally simpler, and scored lower on difficulty indices compared to Brief Hospital Course sections. Additionally, complexity varied significantly by diagnosis: the Eating Disorder cohort uniquely displayed a “readability inversion,” having the simplest instructions but the most complex hospital course summaries. Furthermore, traditional statistical measures showed moderate to high internal consistency. While they demonstrated moderate correlation with MedReadMe, they showed weaker correlation with ClinicalBERT rankings, suggesting that deep learning models capture semantic dimensions of text difficulty distinct from surface-level syntactic complexity. CONCLUSION: Psychiatric documentation practices successfully simplify patient-facing text. However, the observed variability across diagnostic groups and the divergence between traditional and neural readability metrics highlight the necessity of using multipronged evaluation frameworks to accurately assess the accessibility and comprehension of clinical text in the OpenNotes era. </p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13085535/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147390930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hierarchical machine learning models for predicting antenatal care utilisation among Nigerian women: Identifying actionable insights for health policy.","authors":"Augustus Osborne, David B Olawade, Ekomobong Aniefiok Oton, Liwi Martin Odey, Kobloobase Usani","doi":"10.1186/s13040-026-00538-0","DOIUrl":"10.1186/s13040-026-00538-0","url":null,"abstract":"<p><p>BACKGROUND: Ante-natal care (ANC) is critical for improving maternal and neo-natal health outcomes, yet utilisation remains sub-optimal in Nigeria, a country with high maternal mortality. Understanding the patterns and determinants of ANC uptake is essential for designing targeted interventions. This study employs a machine-learning-driven hierarchical modelling framework to predict ANC utilisation among Nigerian women, addressing both initial access and adequacy of care. METHODS: Using a retrospective, cross-sectional design, we analysed the 2024 nationally representative Nigeria Demographic Health Survey data from 13,955 Nigerian women. ANC utilisation was categorized into three classes: no ANC (0 visits), inadequate ANC (1–7 visits), and adequate ANC (≥ 8 visits) per World Health Organisation (WHO) 2016 guidelines. A hierarchical two-stage modelling approach was implemented, with Stage 1 distinguishing any ANC from no ANC using a Light Gradient Boosting Machine (LightGBM) classifier, and Stage 2 differentiating adequate from inadequate ANC among users using an Extreme Gradient Boosting (XGBoost) classifier. Feature engineering captured socioeconomic, demographic, and geographic determinants, while threshold optimisation, calibration, and SHapley Additive exPlanations (SHAP) ensured model robustness and interpretability. Performance was evaluated using precision, recall, F1-score, and area under the receiver operating characteristic curve (AUROC). RESULTS: Of the study population, 27.90% reported no ANC, 54.52% had inadequate ANC, and only 17.58% achieved adequate ANC. Stage 1 achieved an accuracy of 0.77 with high recall (0.87) for any ANC, while Stage 2 reached an accuracy of 0.81 with improved recall (0.80) for adequate ANC after threshold optimisation. The hierarchical framework improved recall of adequate ANC from 0.50 to 0.57 in single-stage models to 0.80 at a policy-oriented operating threshold, highlighting the benefit of sequential modelling. Key predictors included socio-economic status, rural residence, and healthcare access difficulty, with SHAP analyses revealing class-specific effects. CONCLUSIONS: This study demonstrates the potentialof hierarchical machine learning in modelling complex ANC utilisation patterns, providing insights that could inform targeted public health strategies in Nigeria. The low proportion of women achieving adequate ANC offer actionable insights for designing targeted interventions to improve ANC coverage in Nigeria by addressing socioeconomic and geographic barriers. </p>","PeriodicalId":48947,"journal":{"name":"Biodata Mining","volume":" ","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13067523/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147357151","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}