A Hybrid Gradient Boosting and Neural Network Model for Predicting Urban Happiness: Integrating Ensemble Learning with Deep Representation for Enhanced Accuracy

Machine Learning and Knowledge Extraction, Journal Year: 2025, Volume and Issue: 7(1), P. 4 - 4

Published: Jan. 7, 2025

Urban happiness prediction presents a complex challenge, due to the nonlinear and multifaceted relationships among socio-economic, environmental, infrastructural factors. This study introduces an advanced hybrid model combining gradient boosting machine (GBM) neural network (NN) address these complexities. Unlike traditional approaches, this leverages GBM handle structured data features NN extract deeper relationships. The was evaluated against various baseline learning deep models, including random forest, CNN, LSTM, CatBoost, TabNet, using metrics such as RMSE, MAE, R2, MAPE. + achieved superior performance, with lowest RMSE of 0.3332, R2 0.9673, MAPE 7.0082%. also revealed significant insights into urban indicators, 10% improvement in air quality correlating 5% increase happiness. These findings underscore potential models analytics, offering both predictive accuracy actionable for planners.

Language: Английский

---

Hybrid ViT-CapsNet Framework for Brain Tumor Diagnosis Using Biomedical MRI

Biomedical & Pharmacology Journal, Journal Year: 2025, Volume and Issue: 18(December Spl Edition), P. 99 - 119

Published: Jan. 20, 2025

Brain tumor identification through Bio-medical magnetic resonance imaging (MRI) presents a critical challenge in diagnostic imaging, where high accuracy is essential for informed treatment planning. Traditional methods face limitations segmentation precision, leading to increased misdiagnosis risks. This study introduces hybrid deep-learning model integrating Vision Transformer (ViT) and Capsule Network (CapsNet) improve brain classification accuracy. The aims enhance sensitivity specificity categorization. Utilising the BRATS2020 dataset, which comprises 6,000 MRI scans across four classes (meningioma, glioma, pituitary tumor, no tumor), dataset was divided into an 80-20 training-testing split. Data pre-processing included scaling, normalization, feature augmentation robustness. ViT-CapsNet assessed alongside individual ViT CapsNet performance using accuracy, recall, F1-score, AUC-ROC metrics. achieved of 90%, precision recall 89%, F1-score 89.5%, outperforming models. yielded 4-5% improvement types, with notable gains gliomas tumors. Unlike prior methods, achieving 88% our demonstrates superior 90%. approach offers promising solution more accurate detection. Future research could explore refining fusion techniques, advanced interpretability expanding model’s application various clinical environments.

Language: Английский

---

Lung Disease Detection Using Scale-Invariant Weighted Ensemble Neural Architecture

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 83 - 94

Published: Jan. 1, 2025

Language: Английский

---

A Vision Transformer-Based Intelligent System For Brain Tumor Diagnosis

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 642 - 656

Published: Jan. 1, 2025

Language: Английский

---

Explainable deep stacking ensemble model for accurate and transparent brain tumor diagnosis

Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 191, P. 110166 - 110166

Published: April 17, 2025

Early detection of brain tumors in MRI images is vital for improving treatment results. However, deep learning models face challenges like limited dataset diversity, class imbalance, and insufficient interpretability. Most studies rely on small, single-source datasets do not combine different feature extraction techniques better classification. To address these challenges, we propose a robust explainable stacking ensemble model multiclass tumor that combines EfficientNetB0, MobileNetV2, GoogleNet, Multi-level CapsuleNet, using CatBoost as the meta-learner improved aggregation classification accuracy. This approach captures complex characteristics while enhancing robustness The proposed integrates CapsuleNet within framework, utilizing to improve We created two large by merging data from four sources: BraTS, Msoud, Br35H, SARTAJ. tackle applied Borderline-SMOTE augmentation. also utilized methods, along with PCA Gray Wolf Optimization (GWO). Our was validated through confidence interval analysis statistical tests, demonstrating superior performance. Error revealed misclassification trends, assessed computational efficiency regarding inference speed resource usage. achieved 97.81% F1 score 98.75% PR AUC M1, 98.32% 99.34% M2. Moreover, consistently surpassed state-of-the-art CNNs, Vision Transformers, other methods classifying across individual datasets. Finally, developed web-based diagnostic tool enables clinicians interact visualize decision-critical regions scans Explainable Artificial Intelligence (XAI). study connects high-performing AI real clinical applications, providing reliable, scalable, efficient solution

Language: Английский

---

Prediction of Hippocampal Signals in Mice Using a Deep Learning Approach for Neurohybrid Technology Applications

Algorithms, Journal Year: 2024, Volume and Issue: 17(6), P. 252 - 252

Published: June 7, 2024

The increasing growth in knowledge about the functioning of nervous system mammals and humans, as well significant neuromorphic technology developments recent decades, has led to emergence a large number brain–computer interfaces neuroprosthetics for regenerative medicine tasks. Neurotechnologies have traditionally been developed therapeutic purposes help or replace motor, sensory cognitive abilities damaged by injury disease. They also potential memory enhancement. However, there are still no fully neurotechnologies neural capable restoring expanding functions, particular memory, humans. In this regard, search new technologies field restoration functions is an urgent task modern neurophysiology, neurotechnology artificial intelligence. hippocampus important brain structure connected information processing brain. aim paper propose approach based on deep networks prediction hippocampal signals CA1 region received biological input CA3 region. We compare results two widely used architectures: reservoir computing (RC) long short-term (LSTM) networks. proposed study can be viewed first step complex development neurohybrid chip, which allows one restore rodent hippocampus.

Language: Английский

---

ViT-CB: Integrating hybrid Vision Transformer and CatBoost to enhanced brain tumor detection with SHAP

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 100, P. 107027 - 107027

Published: Oct. 24, 2024

Language: Английский

---

ViT-Based Face Diagnosis Images Analysis for Schizophrenia Detection

Brain Sciences, Journal Year: 2024, Volume and Issue: 15(1), P. 30 - 30

Published: Dec. 29, 2024

Computer-aided schizophrenia (SZ) detection methods mainly depend on electroencephalogram and brain magnetic resonance images, which both capture physical signals from patients' brains. These inspection techniques take too much time affect compliance cooperation, while difficult for clinicians to comprehend the principle of decisions. This study proposes a novel method using face diagnosis images based traditional Chinese medicine principles, providing non-invasive, efficient, interpretable alternative SZ detection. An innovative image analysis detection, learns feature representations Vision Transformer (ViT) directly images. It provides features distribution visualization quantitative importance each facial region is proposed supplement interpretation increase efficiency in keeping high accuracy. A benchmarking platform comprising 921 diagnostic 6 benchmark methods, 4 evaluation metrics was established. The experimental results demonstrate that our significantly improves performance with 3-10% accuracy scores. Additionally, it found regions rank descending order according as eyes, mouth, forehead, cheeks, nose, exactly consistent clinical experience. Our fully leverages semantic first-introduced SZ, offering strong interpretability capabilities. not only opens new path but also brings tools concepts research application field mental illness.

Language: Английский

---

Iterative Application of UMAP-Based Algorithms for Fully Synthetic Healthcare Tabular Data Generation

Algorithms, Journal Year: 2024, Volume and Issue: 17(12), P. 591 - 591

Published: Dec. 21, 2024

Building on a previously developed partially synthetic data generation algorithm utilizing visualization techniques, this study extends the novel to generate fully tabular healthcare data. In enhanced form, serves as an alternative conventional methods based Generative Adversarial Networks (GANs) or Variational Autoencoders (VAEs). By iteratively applying original methodology, adapted employs UMAP (Uniform Manifold Approximation and Projection), dimensionality reduction technique, validate generated samples through low-dimensional clustering. This approach has been successfully applied three domains: prostate cancer, breast cardiovascular disease. The have rigorously evaluated for fidelity utility. Results show that UMAP-based outperforms GAN- VAE-based across different scenarios. assessments, it achieved smaller maximum distances between cumulative distribution functions of real attributes. utility evaluations, datasets machine learning model performance, particularly in classification tasks. conclusion, method represents robust solution generating secure, high-quality data, effectively addressing scarcity challenges.

Language: Английский

---

Review on Automated Brain Tumor Segmentation using Advanced Deep Learning Techniques: Enhancing Precision and Clinical Applicability

Published: Dec. 4, 2024

Language: Английский

---