FATIH DURMUS

Mr. FATIH DURMUS

Electrical and Electronics Engineering, Ondokuz Mayıs University, Samsun, Turkey (Türkiye)
orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 June 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
FATIH DURMUSFATIH DURMUS
FATIH DURMUS
SKSERAP KARAGOL
SERAP KARAGOL
Over the last 20 years, lithium-ion batteries have become widely used in many fields due to their advantages such as ease of use and low cost. However, there are concerns about the lifetime and reliability of these batteries. These concerns can be addressed by obtaining accurate capacity and health information. This paper proposes a method to predict the capacity of lithium-ion batteries with high accuracy. Four key features were extracted from current and voltage data obtained during charge and discharge cycles. To enhance prediction accuracy, the Pearson correlation coefficient between these features and battery capacities was analyzed and eliminations were made for some batteries. Using a genetic algorithm (GA), the parameter optimization of Convolutional Neural Network (CNN), Backpropagation (BP), and Recurrent Neural Network (RNN) algorithms was performed. The parameters that provide the best performance were determined in a shorter time using GA, which includes natural selection and genetic processes instead of a trial-and-error method. The study employed five metrics—Mean Square Error (MSE), Root Mean Square Error (RMSE), Normalized Root Mean Square Error (NRMSE), Mean Absolute Error (MAE), and Squared Correlation (R2)—to evaluate prediction accuracy. Predictions based on NASA experimental data were compared with the existing literature, demonstrating superior accuracy. Using 100 training data, 68 data predictions were made with a Root Mean Square Error (RMSE) of 0.1176%. This error rate represents an accuracy level 2.5 times higher than similarly accurate studies in the literature.

6 March 2025 in Diagnostics
Verified
Journal Article
Open access
PDF file
Mehmet MeralMehmet Meral
Mehmet Meral
FÖFerdi Özbilgin
Ferdi Özbilgin
FATIH DURMUSFATIH DURMUS
FATIH DURMUS
Background/Objectives: This paper is significant in highlighting the importance of early and precise diagnosis of Parkinson’s Disease (PD) that affects both motor and non-motor functions to achieve better disease control and patient outcomes. This study seeks to assess the effectiveness of machine learning algorithms optimized to classify PD based on vocal characteristics to serve as a non-invasive and easily accessible diagnostic tool. Methods: This study used a publicly available dataset of vocal samples from 188 people with PD and 64 controls. Acoustic features like baseline characteristics, time-frequency components, Mel Frequency Cepstral Coefficients (MFCCs), and wavelet transform-based metrics were extracted and analyzed. The Chi-Square test was used for feature selection to determine the most important attributes that enhanced the accuracy of the classification. Six different machine learning classifiers, namely SVM, k-NN, DT, NN, Ensemble and Stacking models, were developed and optimized via Bayesian Optimization (BO), Grid Search (GS) and Random Search (RS). Accuracy, precision, recall, F1-score and AUC-ROC were used for evaluation. Results: It has been found that Stacking models, especially those fine-tuned via Grid Search, yielded the best performance with 92.07% accuracy and an F1-score of 0.95. In addition to that, the choice of relevant vocal features, in conjunction with the Chi-Square feature selection method, greatly enhanced the computational efficiency and classification performance. Conclusions: This study highlights the potential of combining advanced feature selection techniques with hyperparameter optimization strategies to enhance machine learning-based PD diagnosis using vocal characteristics. Ensemble models proved particularly effective in handling complex datasets, demonstrating robust diagnostic performance. Future research may focus on deep learning approaches and temporal feature integration to further improve diagnostic accuracy and scalability for clinical applications.

8 November 2021 in Arabian Journal for Science and Engineering
Verified
Journal Article
Closed access
PDF file
FATIH DURMUSFATIH DURMUS
FATIH DURMUS
SKSERAP KARAGOL
SERAP KARAGOL

19 August 2024 in Gümüşhane Üniversitesi Fen Bilimleri Enstitüsü Dergisi
Verified
Journal Article
Closed access
FATIH DURMUSFATIH DURMUS
FATIH DURMUS
FÖFerdi Özbilgin
Ferdi Özbilgin
SKSERAP KARAGOL
SERAP KARAGOL
Radiographic imaging is a crucial tool frequently employed by dentists for initial diagnosis and treatment planning. However, these images often suffer from distortion or inaccuracies due to incorrect exposure settings, making it challenging to identify critical regions such as tooth roots and margins. This study addresses these issues by presenting two innovative methods for tooth segmentation from radiographs, aimed at isolating the tooth regions for better analysis. The first method utilizes fuzzy logic rules to detect edges within the radiographic images. These detected edges are then used as a mask for the Active Contour Method (ACM) to segment the teeth accurately. The second method involves the creation of a Convolutional Neural Network (CNN) for tooth segmentation. The segmentation performance of the CNN is further refined using the ACM, leveraging the initial segmentation as a mask. Both methods demonstrated notable results with varying performance metrics. Specifically, the Fuzzy-Based Active Contour Method achieved precision, recall, and F1 score values of 0.6246, 0.4169, and 0.50, respectively. In contrast, the CNN-Based Active Contour Method calculated accuracy and specificity values of 0.9706 and 0.9872, respectively. These findings indicate that both approaches have distinct strengths in different performance aspects. Our study suggests that these advanced segmentation techniques can significantly enhance the diagnostic capabilities of dental professionals by providing clearer images of tooth structures, aiding in the detection of issues such as root problems, fractures, and wear patterns. Implementing these methods either independently or in combination could lead to more accurate diagnoses and better patient outcomes. Future work could explore the integration of these techniques to leverage their complementary strengths, potentially leading to even greater segmentation accuracy and reliability.

16 June 2026 in Neural Computing and Applications
Verified
Journal Article
Open access
PDF file
FATIH DURMUSFATIH DURMUS
FATIH DURMUS
SKSERAP KARAGOL
SERAP KARAGOL
In this study, a method is proposed to predict the capacity of lithium-ion batteries with high accuracy. Three features, specifically the constant current mode charging time, the discharging time to reach cutoff voltage, and the average discharge voltage, were extracted from the voltage data recorded during the charge and discharge cycles. The Pearson correlation coefficients between these features and the battery capacities were analyzed, showing that the selected features are suitable for all batteries. The parameters of the Convolutional Neural Network (CNN), Back Propagation (BP), and Recurrent Neural Network (RNN) models were optimized using a Genetic Algorithm (GA) with 25%, 50%, and 75% of the training data. Instead of using standard or random values, the GA helped find the best parameter combinations in a shorter time by applying natural selection and genetic processes. The performance of the proposed method was evaluated using five metrics: Mean Square Error (MSE), Coefficient of Determination (R²), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Normalized Root Mean Square Error (NRMSE). The results obtained from the Oxford experimental dataset were compared with other studies in the literature to verify the robustness and superior accuracy of the proposed method.

7 November 2026 in Electronics
Verified
Journal Article
Open access
PDF file
FATIH DURMUSFATIH DURMUS
FATIH DURMUS
FÖFerdi Özbilgin
Ferdi Özbilgin
HCHüseyin CALIK
Hüseyin CALIK
UÇUǧur Çeltekligil
Uǧur Çeltekligil
In this study, 20 methods are evaluated for daily electricity-demand forecasting in Giresun, Türkiye, at one-, two-, and three-day horizons. Consumption-only inputs (S1) are compared with calendar, weather, and event inputs (S2). Forecast-origin temperature persistence and training-only degree-day calibration are used in S2, with origin-day observations assumed available at issuance. Common fitting labels, separate tuning and stopping stages, temporal out-of-fold combination, and condition-specific hyperparameter selection are used over three nonoverlapping outer periods comprising 639 test days. Metrics are calculated from daily seed-mean forecasts, with seed dispersion reported separately. Under S2, a T+1 MAPE of 1.924% and an RMSE of 54.32 MWh are obtained with AS-Pair. The lowest T+2 MAPE (2.709%) is obtained with inverse-MSE blending, while the lowest T+3 MAPE (3.027%) is obtained with CNN-GRU. Lower MAPE is obtained under S2 in 48 of 54 non-naive method–horizon comparisons, with an overall mean reduction of 0.625 percentage points. These results support horizon-specific model selection for daily electricity-demand forecasting in Giresun.