Intelligence-based medicine最新文献

{"title":"Hybrid vision transformer and Xception model for reliable CT-based ovarian neoplasms diagnosis","authors":"Eman Hussein Alshdaifat ,&nbsp;Hasan Gharaibeh ,&nbsp;Amer Mahmoud Sindiani ,&nbsp;Rola Madain ,&nbsp;Asma'a Mohammad Al-Mnayyis ,&nbsp;Hamad Yahia Abu Mhanna ,&nbsp;Rawan Eimad Almahmoud ,&nbsp;Hanan Fawaz Akhdar ,&nbsp;Mohammad Amin ,&nbsp;Ahmad Nasayreh ,&nbsp;Raneem Hamad","doi":"10.1016/j.ibmed.2025.100227","DOIUrl":"10.1016/j.ibmed.2025.100227","url":null,"abstract":"<div><div>Ovarian cancer is a major global health concern, characterized by high mortality rates and a lack of accurate diagnostic methods. Rapid and accurate detection of ovarian cancer is essential to improve patient outcomes and formulate appropriate treatment protocols. Medical imaging methods are essential for identifying ovarian cancer; however, achieving accurate diagnosis remains a challenge. This paper presents a robust methodology for ovarian cancer detection, including the identification and classification of benign and malignant tumors, using the Xception_ViT model. This hybrid approach was chosen because it combines the advantages of traditional CNN-based models (such as Xception) with the capabilities of modern Transformers-based models (such as ViT). This combination allows the model to take advantage of Xception, which extracts features from images. The Vision Transformer (ViT) model is then used to identify connections between diverse visual elements, enhancing the model's understanding of complex components. A Multi-Layer Perceptron (MLP) layer is finally integrated with the proposed model for image classification. The effectiveness of the model is evaluated using three computed tomography (CT) image datasets from King Abdullah University Hospital (KAUH) in Jordan. The first dataset consists of the ovarian cancer computed tomography dataset (KAUH-OCCTD), the second is the benign ovarian tumors dataset (KAUH-BOTD), and the third is the malignant ovarian tumors dataset (KAUH-MOTD). The three datasets collected from 500 women are characterized by their diversity in ovarian tumor classification and are the first of their kind collected in Jordan. The proposed model Xception_ViT achieved an accuracy of 98.09 % in identifying ovarian cancer on the KAUH-OCCTD dataset, and an accuracy of 96.05 % and 98.73 % on the KAUH-BOTD and KAUH-MOTD datasets, respectively, in distinguishing between benign and malignant ovarian tumors. The proposed model outperformed the pre-trained models on all three datasets. The results demonstrate that the proposed model can classify ovarian tumors. This method could also greatly enhance the efficiency of novice radiologists in evaluating ovarian malignancies and assist gynecologists in providing improved treatment alternatives for these individuals.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"11 ","pages":"Article 100227"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465514","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting maternal health risk using PCA-enhanced XGBoost and SMOTE-ENN for improved healthcare outcomes","authors":"Rahmatul Kabir Rasel Sarker ,&nbsp;Sadman Hafij ,&nbsp;Md Adib Yasir ,&nbsp;Md Assaduzzaman ,&nbsp;Md Monir Hossain Shimul ,&nbsp;Md Kamrul Hossain","doi":"10.1016/j.ibmed.2025.100300","DOIUrl":"10.1016/j.ibmed.2025.100300","url":null,"abstract":"<div><h3>Background</h3><div>Maternal health remains a global priority, especially in low-resource settings where timely risk identification is critical. Traditional machine learning models often suffer from poor generalizability, data imbalance, and computational inefficiencies. This study proposes an enhanced predictive model combining SMOTE-ENN data balancing and Principal Component Analysis (PCA) with XGBoost to improve maternal risk classification accuracy using minimal, easily collectible clinical features.</div></div><div><h3>Methods</h3><div>The dataset of 1014 maternal health records comprising seven physiological features was sourced from a public repository. Preprocessing involved standardization, label encoding, and class balancing using SMOTE-ENN. PCA was applied for dimensionality reduction to enhance computational performance and reduce overfitting. Several machine learning classifiers including Decision Tree, Random Forest, LightGBM, Gradient Boosting, and SVM were evaluated, with XGBoost selected as the final model. Performance metrics included accuracy, precision, recall, F1-score, ROC-AUC, and 10-fold cross-validation.</div></div><div><h3>Results</h3><div>The PCA-enhanced XGBoost model achieved the highest accuracy (97.73 %), precision (98 %), recall (98 %), and F1-score (98 %). It outperformed all other models, particularly in identifying high-risk cases with minimal false negatives. Cross-validation confirmed the model's robustness (mean accuracy: 98.39 %), and ROC-AUC scores exceeded 0.998 for all classes, indicating near-perfect classification performance.</div></div><div><h3>Conclusion</h3><div>This study validates a maternal health risk prediction model that is scalable for use in resource-constrained environments and interpretable within the limitations of the selected dimensionality-reduction approach. Its simplicity, high accuracy, and generalizability make it a promising tool for early clinical decision-making and intervention.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"12 ","pages":"Article 100300"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145219043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring the intersection of cochlear implants and artificial intelligence: A mixed-method systematic and scoping review","authors":"Aurenzo Gonçalves Mocelin ,&nbsp;Pedro Angelo Basei de Paula ,&nbsp;Daniel Tiepolo Kochinski ,&nbsp;Thayná Cristina Wiezbicki ,&nbsp;Rogério de Azevedo Hamerschmidt ,&nbsp;Mayara Risnei Watanabe ,&nbsp;Rogério Hamerschmidt","doi":"10.1016/j.ibmed.2025.100296","DOIUrl":"10.1016/j.ibmed.2025.100296","url":null,"abstract":"<div><h3>Objective</h3><div>This study systematically evaluates the role of artificial intelligence (AI) in cochlear implant (CI) technology, focusing on speech enhancement, automated fitting, AI-assisted surgery, predictive modeling, and rehabilitation. The review identifies key advancements, existing limitations, and areas for future development.</div></div><div><h3>Methods</h3><div>Following PRISMA guidelines, we conducted a systematic search across PubMed, IEEE Xplore, Scopus, ScienceDirect, and Embase. We included peer-reviewed primary data studies on AI applications in CIs. The selected studies were categorized into thematic subdomains, such as noise suppression, adaptive programming, AI-driven surgical planning, and telemedicine applications.</div></div><div><h3>Results</h3><div>From an initial pool of 743 records, 129 studies met the eligibility criteria and were included in the final analysis. These studies were categorized into eleven thematic subdomains. The review identified the main application areas and emerging research fronts at the intersection of artificial intelligence and cochlear implant technologies, including speech enhancement, automated fitting, predictive modeling, rehabilitation support, and AI-assisted surgery.</div></div><div><h3>Discussion and conclusion</h3><div>AI is transforming CI technology by improving speech perception, personalization, and surgical precision. However, challenges persist, including computational constraints, data heterogeneity, and the need for large-scale clinical validation. Future research should prioritize energy-efficient AI architectures, regulatory approval pathways, and ethical considerations in automated decision-making. Advancing AI-driven telemedicine solutions can expand CI accessibility, reducing the need for in-person programming. Addressing these challenges will accelerate the development of more adaptive and user-centered CI solutions, ultimately enhancing auditory rehabilitation and quality of life for CI users.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"12 ","pages":"Article 100296"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145264862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Expression of concern for DieT Transformer model with PCA-ADE integration for advanced multi-class brain tumor classification by Mohammad Amin, Khalid M.O. Nahar, et al. [Intell.-Base Med. 11, (2025), 100192, https://doi.org/10.1016/j.ibmed.2024.100192]","authors":"","doi":"","DOIUrl":"","url":null,"abstract":"","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"12 ","pages":"Article 100279"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145733186","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Forecasting pediatric emergency department arrivals: Evaluating the role of exogenous variables using deep learning models","authors":"Egbe-Etu Etu ,&nbsp;Jordan Larot ,&nbsp;Kindness Etu ,&nbsp;Joshua Emakhu ,&nbsp;Sara Masoud ,&nbsp;Imokhai Tenebe ,&nbsp;Gaojian Huang ,&nbsp;Satheesh Gunaga ,&nbsp;Joseph Miller","doi":"10.1016/j.ibmed.2025.100313","DOIUrl":"10.1016/j.ibmed.2025.100313","url":null,"abstract":"<div><h3>Background</h3><div>Forecasting pediatric emergency department (ED) demand remains a critical challenge in healthcare operations. This study aimed to identify exogenous variables influencing pediatric ED visits and evaluate the performance of different forecasting models.</div></div><div><h3>Method</h3><div>Using a retrospective observational design, we analyzed 192,347 pediatric ED visits across nine hospitals in Southeast Michigan between 2017 and 2019. Patient data were aggregated into daily arrival counts and enriched with exogenous variables such as weather, air quality, pollen, calendar, Google search trends, and chief complaints. Feature selection was performed using XGBoost and SHapley Additive exPlanations to identify the most influential predictors. Three forecasting models were developed: a Naïve baseline, Long Short-Term Memory (LSTM), and an attention-based neural network. The models were evaluated across 1-day, 7-day, and 14-day forecasting horizons using mean absolute percentage error (MAPE) and R² metrics.</div></div><div><h3>Results</h3><div>LSTM and attention-based model significantly outperformed the Naïve baseline across all horizons. The LSTM model incorporating calendar data achieved the best 1-day forecast (MAPE: 8.71 %, R²: 0.67). For 7-day forecasts, the attention-based model using chief complaint data performed best (MAPE: 9.18 %, R²: 0.57). At 14 days, the attention-based model without exogenous inputs outperformed most LSTM variants, reflecting superior performance in long-range forecasting. Among exogenous variables, calendar and chief complaint data added the most predictive value, while Google Trends and pollen data introduced noise and diminished model performance.</div></div><div><h3>Conclusion</h3><div>Combining deep learning architectures with selected external data improves pediatric ED arrival forecasting. From an operational perspective, such forecasts can support more efficient staffing, reduce wait times, and mitigate ED crowding.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"12 ","pages":"Article 100313"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145519506","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting the prevalence of cardiovascular diseases using machine learning algorithms","authors":"Bernada E. Sianga ,&nbsp;Maurice C. Mbago ,&nbsp;Amina S. Msengwa","doi":"10.1016/j.ibmed.2025.100199","DOIUrl":"10.1016/j.ibmed.2025.100199","url":null,"abstract":"<div><div>Cardiovascular Diseases (CVDs) are the major cause of morbidity, disability, and mortality worldwide and are the most life-threatening diseases. Early detection and appropriate action can significantly reduce the effects and complications of CVD. Prediction of the likelihood that an individual can develop CVD adverse outcomes is essential. Machine learning methods are used to predict the risk of CVD incidences. Optimal model parameters were obtained using the grid search and randomized search methods. A hyperparameter tuning method with the highest accuracy was used to find the optimal parameters for the six algorithms used in this study. Two experiments were deployed: the first was training and testing the CVD dataset using hyperparameterized ML algorithms excluding geographical features, and the second included geographical features. The geographical features are air humidity, temperature and education status of a location. The performances of the two experiments were compared using classification metrics. The findings revealed that the performance of the second experiment outperformed the first experiment. XGBoost achieved the highest accuracy of 95.24 %, followed by the decision tree 93.87 % and support vector machine 92.87 % when geographical features were included (second experiment). Including geographical risk factors in predicting CVD is crucial as they contribute to the probability of developing CVD incidences.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"11 ","pages":"Article 100199"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174330","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A neonatal sepsis prediction algorithm using electronic medical record data from Mbarara Regional Referral Hospital","authors":"Peace Ezeobi Dennis ,&nbsp;Angella Musiimenta ,&nbsp;William Wasswa ,&nbsp;Stella Kyoyagala","doi":"10.1016/j.ibmed.2025.100198","DOIUrl":"10.1016/j.ibmed.2025.100198","url":null,"abstract":"<div><h3>Introduction</h3><div>Neonatal sepsis is a global challenge that contributes significantly to neonatal morbidity and mortality. The current diagnostic methods depend on conventional culture methods, a procedure that takes time and leads to delays in making timely treatment decisions. This study proposes a machine learning algorithm utilizing electronic medical record (EMR) data from Mbarara Regional Referral Hospital (MRRH) to enhance early detection and treatment of neonatal sepsis.</div></div><div><h3>Methods</h3><div>We performed a retrospective study on a dataset of neonates hospitalized for at least 48 h in the Neonatal Intensive Care Unit (NICU) at MRRH between October 2015 to September 2019 who received at least one sepsis evaluation. 482 records of neonates met the inclusion criteria and the dataset comprises 38 neonatal sepsis screening parameters. The study considered two outcomes for sepsis evaluations: culture-positive if a blood culture was positive, and clinically positive if cultures were negative but antibiotics were administered for at least 120 h. We implemented k-fold cross-validation with k set to 10 to guarantee robust training and testing of the models. Seven machine learning models were trained to classify inputs as sepsis positive or negative, and their performance was compared with physician diagnoses.</div></div><div><h3>Results</h3><div>The results of this study show that the proposed algorithm, combining maternal risk factors, neonatal clinical signs, and laboratory tests (the algorithm demonstrated a sensitivity and specificity of at least 95 %) outperformed the physician diagnosis (Sensitivity = 89 %, Specificity = 11 %). SVM model with radial basis function, polynomial kernels, and DT model (with the highest AUROC values of 98 %) performed better than the other models.</div></div><div><h3>Conclusions</h3><div>The study shows that the combination of maternal risk factors, neonatal clinical signs, and laboratory tests can help improve the prediction of neonatal sepsis. Further research is warranted to assess the potential performance improvements and clinical efficacy in a prospective trial.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"11 ","pages":"Article 100198"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning-based approach to diagnose lung cancer using CT-scan images","authors":"Mohammad Q. Shatnawi,&nbsp;Qusai Abuein,&nbsp;Romesaa Al-Quraan","doi":"10.1016/j.ibmed.2024.100188","DOIUrl":"10.1016/j.ibmed.2024.100188","url":null,"abstract":"<div><div>The work in this research focuses on the automatic classification and prediction of lung cancer using computed tomography (CT) scans, employing Deep Learning (DL) strategies, specifically Enhanced Convolutional Neural Networks (CNNs), to enable rapid and accurate image analysis. This research designed and developed pre-trained models, including ConvNeXtSmall, VGG16, ResNet50, InceptionV3, and EfficientNetB0, to classify lung cancer. The dataset was divided into four classes, consisting of 338 images of adenocarcinoma, 187 images of large cell carcinoma, 260 images of squamous cell carcinoma, and 215 normal images. Notably, The Enhanced CNN model achieved an unprecedented testing accuracy of 100 %, outperforming all other models, which included ConvNeXt at 87 %, VGG16 at 99 %, ResNet50 at 94.5 %, InceptionV3 at 76.9 %, and EfficientNetB0 at 97.9 %. The study of this research is considered the first one that hits 100 % testing accuracy with an Enhanced CNN, demonstrating significant advancements in lung cancer detection through the application of sophisticated image enhancement techniques and innovative model architectures. This highlights the potential of Enhanced CNN models in transforming lung cancer diagnostics and emphasizes the importance of integrating advanced image processing techniques into clinical practice.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"11 ","pages":"Article 100188"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143174358","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Attention-driven graph-based machine learning for non-invasive diagnosis of NAFLD","authors":"Ekta Srivastava ,&nbsp;Sarath Mohan ,&nbsp;Tapan Kumar Gandhi ,&nbsp;Ashok Kumar Choudhury ,&nbsp;Sandeep Kumar","doi":"10.1016/j.ibmed.2025.100288","DOIUrl":"10.1016/j.ibmed.2025.100288","url":null,"abstract":"<div><div>An estimated 25%–30% of the global population is affected by non-alcoholic fatty liver disease (NAFLD), a silent yet progressive condition that can advance from simple steatosis to severe stages like non-alcoholic steatohepatitis (NASH), fibrosis, and cirrhosis, significantly heightening the risk of liver cancer. Currently, the gold-standard method for staging NAFLD is liver biopsy, an invasive procedure with risks such as bleeding, infection, and sampling error. Due to its high cost and impracticality for routine monitoring, there is a critical need for reliable, non-invasive diagnostic tools capable of effectively identifying NAFLD stages. We developed a graph-based framework in which each patient is represented as a node in a similarity network. Edges are formed via k-nearest neighbors (KNN) on standardized clinical and biochemical features, with missing values imputed by KNN to preserve biologically plausible variability. A two-layer Graph Attention Network (GAT) then learns edge-specific attention weights to focus on the most informative inter-patient relationships. Tested on a proprietary ILBS cohort (n = 622), our model achieved 75.2% accuracy (AUC = 0.768; F1 = 0.752), an 11% absolute improvement over Support Vector Machines and Random Forests, and demonstrated robustness in 10-fold cross-validation and adversarial noise tests. On a separate public dataset (n = 80) spanning lipidomic, glycomic, fatty acid, and hormone panels, it exceeded 99% accuracy (AUC <span><math><mo>&gt;</mo></math></span> 0.99). Attention-based explanations further highlighted key patient similarities driving each prediction. These findings suggest that attention-driven graph learning can clearly improve non-invasive NAFLD staging, enabling early detection and supporting personalized disease monitoring in diverse clinical settings.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"12 ","pages":"Article 100288"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144912336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Privacy-aware and interpretable deep learning framework for dental caries classification","authors":"Jashvant Kumar ,&nbsp;Khaled Mohamad Almustafa ,&nbsp;Rand Madanat ,&nbsp;Akhilesh Kumar Sharma ,&nbsp;Muhammed Sutcu ,&nbsp;Juliano Katrib","doi":"10.1016/j.ibmed.2025.100294","DOIUrl":"10.1016/j.ibmed.2025.100294","url":null,"abstract":"<div><div>Dental caries remains one of the most prevalent and persistent chronic diseases globally, affecting individuals across all age groups and posing a significant burden on public health systems. Early detection is critical to prevent the progression of tooth decay, reduce treatment complexity, and improve long-term oral health outcomes. In response to these clinical demands, this study presents a comprehensive, privacy-aware, and interpretable deep learning framework for the automated classification of dental caries from X-ray images. The approach addresses the issues of class imbalance, low Resolution image and privacy preserved patient's medical images.The framework is structured into three progressive phases that incorporate supervised learning through Convolutional Neural Networks (CNN), ResNet-18, and DenseNet; unsupervised clustering using Principal Component Analysis (PCA); and a decentralized federated learning strategy to ensure secure model training across distributed datasets. The experimental dataset consists of 957 labelled dental radiographs, including 174 healthy and 783 carious cases, emphasizing the issue of class imbalance. Initial baseline models achieved an accuracy of 84 %, which improved to 96 % following strategic data augmentation and class balancing interventions. PCA-based clustering visualizations revealed well-separated clusters (Silhouette Score: 0.6660), confirming the discriminative power of the selected features. Meanwhile, the federated learning implementation preserved data confidentiality without sacrificing performance, reinforcing the model's suitability for real-world clinical deployment. Collectively, these findings validate the framework's robustness, interpretability, and adaptability, offering a scalable and ethically aligned solution for AI-driven dental diagnostics in modern healthcare systems.</div></div>","PeriodicalId":73399,"journal":{"name":"Intelligence-based medicine","volume":"12 ","pages":"Article 100294"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144912406","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}