Ruya Yilmaz
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ruya Yilmaz has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Ahmed Lateef Salih Al-Karawi
Department of Computer Engineering, Faculty of Engineering, Istanbul Aydin University, Istanbul 34295, Turkey
2 shared publications - Hayder Mohammedqasim
Istanbul Aydın University, Istanbul, Türkiye, Turkey (Türkiye)
2 shared publications - Ahmed Qays Kamil
Istanbul Aydın University, Istanbul, Türkiye, Turkey (Türkiye)
1 shared publication - Qays K.N. Alkhlidy
General Commission for Groundwater, Ministry of Education, Baghdad, Iraq, Iraq
1 shared publication
Latest Publications
- Open Access
- 22 April 2026
- Published by MDPI AG in Preprints.org
An Integrated EWQI-Guided Machine Learning Framework for Groundwater Quality: Regression and Multi-Class Classification under Iraqi Standards
Groundwater is a major source of water supply in Diyala Governorate, Iraq. This study proposes an integrated framework for groundwater quality assessment and forecasting by combining the Entropy Weighted Water Quality Index (EWQI) with machine learning algorithms. A total of 855 groundwater samples were collected, of which 853 were retained after preprocessing. The EWQI was calculated using ten hydrochemical parameters for each sample, namely pH, total dissolved solids (TDS), K+, Na+, Mg2+, Ca2+, Cl−, SO42−, HCO3−, and NO3−. Based on the Iraqi Standard Drinking Water Limits (IQS), EWQI values were categorized into five water quality classes: Excellent, Good, Medium, Poor, and Extremely Poor. For predictive analysis, both regression and classification models were developed using support vector machine (SVM), random forest (RF), backpropagation multilayer perceptron (BP-MLP), and one-dimensional convolutional neural network (1D-CNN) algorithms under three feature scenarios: (1) hydrochemical variables only, (2) hydrochemical variables combined with well properties, and (3) hydrochemical variables, well properties, and spatial coordinates. Model hyperparameters were optimized using grid search with 10-fold cross-validation on the training set, and model performance was evaluated using an 80/20 stratified split. Among the tested models, SVM demonstrated the best performance for EWQI prediction in the first scenario (hydrochemical variables only), achieving RMSE = 4.14, MAE = 0.93, and R2 = 0.999. It also produced the highest classification performance, with an accuracy of 0.971 and a macro-F1 score of 0.973. Meanwhile, RF showed consistently strong and balanced classification performance across all scenarios, reaching a macro-F1 score of 0.961. These findings highlight the effectiveness of integrating EWQI with machine learning techniques for reliable groundwater quality assessment and prediction, thereby supporting sustainable water resource management in Diyala Governorate.
- 22 April 2026
- Published by MDPI AG in Preprints.org
Open AccessAn Integrated EWQI-Guided Machine Learning Framework for Groundwater Quality: Regression and Multi-Class Classification under Iraqi Standards
Groundwater is a major source of water supply in Diyala Governorate, Iraq. This study proposes an integrated framework for groundwater quality assessment and forecasting by combining the Entropy Weighted Water Quality Index (EWQI) with machine learning algorithms. A total of 855 groundwater samples were collected, of which 853 were retained after preprocessing. The EWQI was calculated using ten hydrochemical parameters for each sample, namely pH, total dissolved solids (TDS), K+, Na+, Mg2+, Ca2+, Cl−, SO42−, HCO3−, and NO3−. Based on the Iraqi Standard Drinking Water Limits (IQS), EWQI values were categorized into five water quality classes: Excellent, Good, Medium, Poor, and Extremely Poor. For predictive analysis, both regression and classification models were developed using support vector machine (SVM), random forest (RF), backpropagation multilayer perceptron (BP-MLP), and one-dimensional convolutional neural network (1D-CNN) algorithms under three feature scenarios: (1) hydrochemical variables only, (2) hydrochemical variables combined with well properties, and (3) hydrochemical variables, well properties, and spatial coordinates. Model hyperparameters were optimized using grid search with 10-fold cross-validation on the training set, and model performance was evaluated using an 80/20 stratified split. Among the tested models, SVM demonstrated the best performance for EWQI prediction in the first scenario (hydrochemical variables only), achieving RMSE = 4.14, MAE = 0.93, and R2 = 0.999. It also produced the highest classification performance, with an accuracy of 0.971 and a macro-F1 score of 0.973. Meanwhile, RF showed consistently strong and balanced classification performance across all scenarios, reaching a macro-F1 score of 0.961. These findings highlight the effectiveness of integrating EWQI with machine learning techniques for reliable groundwater quality assessment and prediction, thereby supporting sustainable water resource management in Diyala Governorate.
- Open Access
- 16 April 2026
- Published by MDPI AG in Preprints.org
Federated Privacy-Preserving Multi-Modal Deep Learning for Breast Cancer Diagnosis: A Physics-Aware Approach
Breast cancer remains a leading cause of cancer-related mortality among women globally. This study makes one focused primary contribution: a formalized, physics-grounded preprocessing-to-fusion pipeline for multi-modal breast cancer classification that is rigorously validated under both centralized and federated learning conditions. Patient-wise stratified 5-fold cross-validation was applied across Ultrasound (BUSI, n=780), Dynamic Contrast-Enhanced MRI (DUKE, n=922), and Mammography (CBIS-DDSM, n=400). Per-modality models achieved 92.50±1.2%, 90.63±1.5%, and 92.00±1.3% accuracy (McNemar’s p<0.05 vs. baselines). Weighted late-fusion achieved 93.10±1.1% (p=0.031 vs. best individual modality). A five-algorithm FL comparison (FedAvg, FedProx, SCAFFOLD, FedNova, FP16-FedAvg) under IID and non-IID (Dirichlet α=0.5) conditions is provided with per-round training time, communication time, per-round latency, and cumulative bandwidth. FP16 transmission reduced bandwidth from 8.14 GB to 1.23 GB (−84.9%, p=0.74 vs. FP32). SCAFFOLD achieved the best non-IID accuracy (90.50%). All design choices are validated by ablation experiments with McNemar’s test and Cohen’s h effect sizes.
- 16 April 2026
- Published by MDPI AG in Preprints.org
Open AccessFederated Privacy-Preserving Multi-Modal Deep Learning for Breast Cancer Diagnosis: A Physics-Aware Approach
Breast cancer remains a leading cause of cancer-related mortality among women globally. This study makes one focused primary contribution: a formalized, physics-grounded preprocessing-to-fusion pipeline for multi-modal breast cancer classification that is rigorously validated under both centralized and federated learning conditions. Patient-wise stratified 5-fold cross-validation was applied across Ultrasound (BUSI, n=780), Dynamic Contrast-Enhanced MRI (DUKE, n=922), and Mammography (CBIS-DDSM, n=400). Per-modality models achieved 92.50±1.2%, 90.63±1.5%, and 92.00±1.3% accuracy (McNemar’s p<0.05 vs. baselines). Weighted late-fusion achieved 93.10±1.1% (p=0.031 vs. best individual modality). A five-algorithm FL comparison (FedAvg, FedProx, SCAFFOLD, FedNova, FP16-FedAvg) under IID and non-IID (Dirichlet α=0.5) conditions is provided with per-round training time, communication time, per-round latency, and cumulative bandwidth. FP16 transmission reduced bandwidth from 8.14 GB to 1.23 GB (−84.9%, p=0.74 vs. FP32). SCAFFOLD achieved the best non-IID accuracy (90.50%). All design choices are validated by ablation experiments with McNemar’s test and Cohen’s h effect sizes.