Researching public health datasets in the era of deep learning: a systematic literature review

Health Informatics Journal, Journal Year: 2025, Volume and Issue: 31(1)

Published: Jan. 1, 2025

Objective: Explore deep learning applications in predictive analytics for public health data, identify challenges and trends, then understand the current landscape. Materials Methods: A systematic literature review was conducted June 2023 to search articles on data context of learning, published from inception medical computer science databases through 2023. The focused diverse datasets, abstracting applications, challenges, advancements learning. Results: 2004 were reviewed, identifying 14 disease categories. Observed trends include explainable-AI, patient embedding integrating different sources employing models informatics. Noted technical reproducibility handling sensitive data. Discussion: There has been a notable surge publications since 2015. Consistent continue be applied across Despite wide standard approach still does not exist addressing outstanding issues this field. Conclusion: Guidelines are needed applying improve FAIRness, efficiency, transparency, comparability, interoperability research. Interdisciplinary collaboration among scientists, experts, policymakers is harness full potential

Language: Английский

---

Optimizing brain tumor classification through feature selection and hyperparameter tuning in machine learning models

Informatics in Medicine Unlocked, Journal Year: 2023, Volume and Issue: 43, P. 101414 - 101414

Published: Jan. 1, 2023

Accurately classifying brain tumors using images is extremely important for prognosis and treatment planning. In this study, we have developed an optimized approach machine learning techniques to classify tumors. Our method involves preprocessing the images, extracting features, selecting most significant ones, tuning model parameters. We utilized filtering, morphological opening, normalization enhance image quality reduce noise. extracted 17 features that capture characteristics of identify seven distinguishing through importance analysis. By employing a range models such as Random Forest, Support Vector Machines, Extreme Gradient Boosting, K Nearest Neighbors, Categorical Extra Trees, Naive Bayes, achieve accuracy 98.0 % after thorough hyperparameter optimization. This research highlights impact feature selection process, along with tuning, on maximizing classification performance. provides framework enables diagnosis enhanced clinical decision-making patient care.

Language: Английский

---

Enhancing Brain Disease Diagnosis with XAI: A Review of Recent Studies

ACM Transactions on Computing for Healthcare, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 9, 2025

The area of eXplainable Artificial Intelligence (XAI) has shown remarkable progress in the past few years, with aim enhancing transparency and interpretability machine learning (ML) deep (DL) models. This review paper presents an in-depth current state-of-the-art XAI techniques applied to diagnosis brain diseases. challenges encountered by traditional ML DL models within this domain are thoroughly examined, emphasising pivotal role providing these Furthermore, a comprehensive survey methodologies used for making diagnoses various disorders. Recent studies utilising diagnosing range illnesses, including Alzheimer, tumours, dementia, Parkinson, multiple sclerosis, autism, epilepsy, stroke, critically reviewed. Finally, limitations inherent discussed, along prospective avenues future research. key goal study is provide researchers roadmap that shows potential improving algorithms diseases, while also delineating require concerted research efforts.

Language: Английский

---

Enhancing Transparency and Trust in Brain Tumor Diagnosis: An In-Depth Analysis of Deep Learning and Explainable AI Techniques

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 30, 2025

Brain tumors pose significant health risks due to their high mortality rates and challenges in early diagnosis. Advances medical imaging, particularly MRI, combined with artificial intelligence (AI), have revolutionized tumor detection, segmentation, classification. Despite the accuracy of models such as Convolutional Neural Networks (CNNs) Vision Transformers (ViTs), clinical adoption is hampered by a lack interpretability. This study provides comprehensive analysis machine learning, deep explainable AI (XAI) techniques brain diagnosis, emphasizing strengths, limitations, potential improve transparency trust. By reviewing 53 peer-reviewed articles published between 2017 2024, we assess current state research, identify gaps, provide practical recommendations for clinicians, regulators, developers. The findings reveal that while XAI techniques, Grad-CAM, SHAP, LIME, significantly enhance model interpretability, remain terms generalizability, computational complexity, dataset quality. Future research should focus on addressing these limitations fully realize diagnostics.

Language: Английский

---

Navigating Explainable Ai in Healthcare: A Taxonomic Guide for Clinical Decision Support Systems

Published: Jan. 1, 2025

Language: Английский

---

Extracting Clinical Relationships from Discharge Summaries of Supra Sellar Lesion Patients using Gemini LLM

Procedia Computer Science, Journal Year: 2025, Volume and Issue: 258, P. 2391 - 2404

Published: Jan. 1, 2025

Language: Английский

---

Multi-Class Prediction of Suicide Behavior of Adolescents Using Machine Learning Approach

2022 13th International Conference on Computing Communication and Networking Technologies (ICCCNT), Journal Year: 2023, Volume and Issue: unknown

Published: July 6, 2023

Suicide, a leading global cause of mortality, is highly debated topic. Teenagers face increased psychological stress and self-doubt, making them more vulnerable to suicide. The study aims predict suicide rates examine contributing factors using dataset from WHO's Global School-based Student Health Survey (GSHS). being continuous-valued, regression algorithms like Linear Regression, Random Forest, Ridge KNN Regression were applied. target variable was discretized, it classification problem. Subsequently, such as Naïve Bayes, AdaBoost, XGBoost, Decision Tree, Logistic SVM, KNN, utilized. Gradient Boost, Bagging models show accurate predictions with test RMSE approximately 0. Classification achieved F1-scores ranging 0.68 0.87, Bayes attaining the highest score outperforming other algorithms. methods excel over on continuous-valued features, rendering optimal choice for this study.

Language: Английский

---

EXplainable Artificial Intelligence (XAI) for MRI brain tumor diagnosis: A survey

Published: Oct. 3, 2023

The results of the Deep Learning (DL) are indisputable in different fields and particular that medical diagnosis. black box nature this tool has left doctors very cautious with regard to its estimates. eXplainable Artificial Intelligence (XAI) recently seemed lift challenge by providing explanations DL Several works published literature offering explanatory methods. We interested survey present an overview on application XAI Learning-based Magnetic Resonance Imaging (MRI) image analysis for Brain Tumor (BT) In survey, we divide these methods into four groups, group intrinsic three groups post-hoc which activation based, gradientr based perturbation These tools improved confidence brain tumor

Language: Английский

---

Stacking: An ensemble learning approach to predict student performance in PISA 2022

Education and Information Technologies, Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 29, 2024

Language: Английский

---

Explainable AI Insights into Skin Cancer Detection: A Comparative Study of CNN, DenseNet, and ResNet

2022 IEEE 7th International conference for Convergence in Technology (I2CT), Journal Year: 2024, Volume and Issue: unknown

Published: April 5, 2024

Skin cancer is a dangerous and widespread conditionthat requires early accurate detection for effective treatment. Recent advancements in deep learning have demonstrated promise the of skin from image datasets. This research aims to analyze effectiveness different models detecting cancer, including DenseNet, CNN, ResNet. study evaluates metrics like accuracy, precision, recall, F1-score identifying cancer. Additionally, this investigates important features images that lead model prediction using Explainable AI - LIME SHAP. The ultimate aim discover clever methods early. helps patients get treatment quickly when it matters most.

Language: Английский

---