One-hot constrained symmetric nonnegative matrix factorization for image clustering

Pattern Recognition, Journal Year: 2025, Volume and Issue: 162, P. 111427 - 111427

Published: Feb. 7, 2025

Language: Английский

---

Identifying optimal technique of reducing dimensionality of scour influencing hydraulic parameters apply SWOT analysis

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 126829 - 126829

Published: Feb. 1, 2025

Language: Английский

---

An autoencoder-like deep NMF representation learning algorithm for clustering

Knowledge-Based Systems, Journal Year: 2024, Volume and Issue: 305, P. 112597 - 112597

Published: Oct. 10, 2024

Language: Английский

---

A collaborative filtering recommender systems: Survey

Neurocomputing, Journal Year: 2024, Volume and Issue: unknown, P. 128718 - 128718

Published: Nov. 1, 2024

Language: Английский

---

When graph neural networks meet deep nonnegative matrix factorization: An encoder and decoder-like method for community detection

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 126676 - 126676

Published: Feb. 1, 2025

Language: Английский

---

Deep Supervised Cone-Based Nonnegative Matrix Factorization in Image Pattern Space

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 49 - 60

Published: Jan. 1, 2025

Language: Английский

---

Analysis of deep non-smooth symmetric nonnegative matrix factorization on hierarchical clustering

Applied Intelligence, Journal Year: 2025, Volume and Issue: 55(6)

Published: Feb. 24, 2025

Language: Английский

---

Evaluation of stroke sequelae and rehabilitation effect on brain tumor by neuroimaging technique: A comparative study

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(2), P. e0317193 - e0317193

Published: Feb. 24, 2025

This study aims at the limitations of traditional methods in evaluation stroke sequelae and rehabilitation effect monitoring, especially for accurate identification tracking brain injury areas. To overcome these challenges, we introduce an advanced neuroimaging technology based on deep learning, SWI-BITR-UNet model. model, introduced as novel Machine Learning (ML) combines SWIN Transformer’s local receptive field shift mechanism, effective feature fusion strategy U-Net architecture, aiming to improve accuracy lesion region segmentation multimodal MRI scans. Through application a 3-D CNN encoder decoder, well integration CBAM attention module jump connection, model can finely capture refine features, achieve level comparable that manual by experts. introduces 3D encoder-decoder architecture specifically designed enhance processing capabilities medical imaging data. The development utilizes ADAM optimization algorithm facilitate training process. Bra2020 dataset is utilized assess proposed learning neural network. By employing skip connections, effectively integrates high-resolution features from with up-sampling thereby increasing model’s sensitivity spatial characteristics. both testing phases, SWI-BITR-Unet trained using reliable datasets evaluated through comprehensive array statistical metrics, including Recall (Rec), Precision (Pre), F1 test score, Kappa Coefficient (KC), mean Intersection over Union (mIoU), Receiver Operating Characteristic-Area Under Curve (ROC-AUC). Furthermore, various machine models, such Random Forest (RF), Support Vector (SVM), Extreme Gradient Boosting (XGBoost), Categorical (CatBoost), Adaptive (AdaBoost), K-Nearest Neighbor (KNN), have been employed analyze tumor progression brain, performance characterized Hausdorff distance. In From ML was more than other models. Subsequently, regarding DICE coefficient values, maps (annotation distributions) generated models indicated models’s capability autonomously delineate areas core (TC) enhancing (ET). Moreover, efficacy demonstrated superiority existing research field. computational efficiency ability handle long-distance dependencies make it particularly suitable applications clinical Settings. results showed SNA-BITR-UNet not only identify monitor subtle changes area, but also provided new efficient tool process, providing scientific basis developing personalized plans.

Language: Английский

---

Representation Auto-fused NMF based Hierarchical Clustering

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127560 - 127560

Published: April 1, 2025

Language: Английский

---

High-Order Aligned Deep Complementary and View-Specific Similarity Graphs for Unsupervised Multi-View Feature Selection

Published: Jan. 1, 2025

Language: Английский

---