Fatih ABUT

Dr. Fatih ABUT

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
30 April 2021 in Balkan Journal of Electrical and Computer Engineering
Verified
Journal Article
Open access
PDF file
MAMohammed Abdeh
Mohammed Abdeh
FAFatih ABUT
Fatih ABUT
MAMehmet Fatih Akay
Mehmet Fatih Akay
Human assisted search and rescue (SAR) robots are increasingly being used in zones of natural disasters, industrial accidents, and civil wars. Due to complex terrains, obstacles, and uncertainties in time availability, there is a need for these robots to have a certain level of autonomy to act independently for approaching certain SAR tasks. One of these tasks is autonomous navigation. Previous approaches to develop autonomous or semi-autonomous SAR navigating robots use heuristics-based methods. These algorithms, however, require environment-related prior knowledge and enough sensing capabilities, which are hard to maintain due to restrictions of size and weight in highly unstructured environments such as collapsed buildings. This study approaches the problem of autonomous navigation using a modified version of the Deep Q-Network algorithm. Unlike the classical usage of the entire game screen images to train the agent, our approach uses only the images captured by the agent's low-resolution camera to train the agent for navigating through an arena avoiding obstacles and to reach a victim. This approach is a much more relevant way of decision making in complex, uncertain contexts; since in real-world SAR scenarios, it is almost impossible to have the area's full information to be used by SAR teams. We simulated a SAR scenario, which consists of an arena full of randomly generated obstacles, a victim, and an autonomous SAR robot. The simulation results show that the agent was able to reach the victim in 56% of the evaluation episodes after 400 episodes of training.

25 September 2020 in Turkish Journal of Electrical Engineering and Computer Sciences
Verified
Journal Article
Closed access
ADAlper Kamil Demir
Alper Kamil Demir
FAFatih ABUT
Fatih ABUT

9 May 2025 in Electronics
Verified
Journal Article
Open access
PDF file
MÇMehmet Serhan Çiftlikçi
Mehmet Serhan Çiftlikçi
Yusuf ÇakmakYusuf Çakmak
Yusuf Çakmak
TKTOLGA AHMET KALAYCI
TOLGA AHMET KALAYCI
FAFatih ABUT
Fatih ABUT
MAMehmet Fatih Akay
Mehmet Fatih Akay
MKMehmet Kızıldağ
Mehmet Kızıldağ
In the rapidly evolving field of e-commerce, precise and efficient attribute extraction from product descriptions is crucial for enhancing search functionality, improving customer experience, and streamlining the listing process for sellers. This study proposes a large language model (LLM)-based approach for automated attribute extraction on Trendyol’s e-commerce platform. For comparison purposes, a deep learning (DL) model is also developed, leveraging a transformer-based architecture to efficiently identify explicit attributes. In contrast, the LLM, built on the Mistral architecture, demonstrates superior contextual understanding, enabling the extraction of both explicit and implicit attributes from unstructured text. The models are evaluated on an extensive dataset derived from Trendyol’s Turkish-language product catalog, using performance metrics such as precision, recall, and F1-score. Results indicate that the proposed LLM outperforms the DL model across most metrics, demonstrating superiority not only in direct single-model comparisons but also in average performance across all evaluated categories. This advantage is particularly evident in handling complex linguistic structures and diverse product descriptions. The system has been integrated into Trendyol’s platform with a scalable backend infrastructure, employing Kubernetes and Nvidia Triton Inference Server for efficient bulk processing and real-time attribute suggestions during the product listing process. This study not only advances attribute extraction for Turkish-language e-commerce but also provides a scalable and efficient NLP-based solution applicable to large-scale marketplaces. The findings offer critical insights into the trade-offs between accuracy and computational efficiency in large-scale multilingual NLP applications, contributing to the broader field of automated product classification and information retrieval in e-commerce ecosystems.

2 November 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MKMehmet Kızıldağ
Mehmet Kızıldağ
FAFatih ABUT
Fatih ABUT
MAMehmet Fatih Akay
Mehmet Fatih Akay
The System Marginal Price (SMP) is the cost of the last unit of electricity supplied to the grid, reflecting the supply–demand equilibrium and serving as a key indicator of market conditions. Accurate SMP forecasting is essential for ensuring market stability and economic efficiency. This study addresses the challenges of SMP prediction in Turkey by proposing a comprehensive forecasting framework that integrates machine learning, deep learning, and statistical models. Advanced feature selection techniques, such as Minimum Redundancy Maximum Relevance (mRMR) and Maximum Likelihood Feature Selector (MLFS), are employed to refine model inputs. The framework incorporates time series methods like Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), and Convolutional LSTM (ConvLSTM) to capture complex temporal patterns, alongside models such as Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Extreme Learning Machine (ELM) for modeling non-linear relationships. Model performance was evaluated using the Mean Absolute Percentage Error (MAPE) across regular weekdays, weekends, and public holidays. XGBoost combined with MLFS consistently achieved the lowest MAPE values, demonstrating exceptional accuracy and robustness. Among all of the models, XGBoost combined with MLFS consistently achieved the lowest MAPE values, demonstrating superior accuracy and robustness. The results highlight the inadequacy of traditional models like ARIMA and SARIMA in capturing non-linear and highly volatile patterns, reinforcing the necessity of using advanced techniques for effective SMP forecasting. Overall, this study presents a novel and comprehensive approach tailored for complex electricity markets, significantly enhancing predictive reliability by incorporating economic indicators and sophisticated feature selection methods.

30 November 2022 in Kocaeli Journal of Science and Engineering
Verified
Journal Article
Closed access
HÖH. Onur Özcan
H. Onur Özcan
IÇİsMail Çolak
İsMail Çolak
SESelin Erimhan
Selin Erimhan
VGVedat Güneş
Vedat Güneş
FAFatih ABUT
Fatih ABUT
MAMehmet Fatih Akay
Mehmet Fatih Akay
Fraud detection identifies suspicious activities, false pretenses, wrongful or criminal deception intended to result in financial gain. Fraud is rare, well thought, effortful, and deceiving throughout claims. Detecting fraudulent claims is essential for the insurance industry. Therefore, most insurance companies must devote time and budget to fraud detection. Fraud detection can be divided into two categories; the main and most common type of fraud is individual fraud. Individual frauds can appear in many kinds of forms. For example, damage to an asset might be occurred before issuing a policy and be reported after. The second category is organized fraud which is much rarer and harder to detect than individual fraud. Especially motor insurance fraud is commonly attempted by organized crime rings. Counterparties involved in fraudulent claims change frequently, and changes make fraud detection difficult. According to Insurance Information and Monitoring Center findings, the fraudulent claim payment ratio is 10 to 30 %, and the detection success rate for an individual is at 1.4 to 5%. At the same time, the annual fraud cost is at 200 to 300 $ million. This study proposes a fraud detection platform called SOBE, which assists fraud departments’ claim inquiry more easily and shorter than manual investigation made by employees. At its core, SOBE uses a rule engine approach. In order to support the rule engine, there is also a machine learning algorithm for fraud detection. In addition, the SNA module detects interconnected fraud counterparts among claim files. Consequently, the SOBE fraud detection platform allows Anadolu Sigorta to prevent improper payments from claiming participants. SOBE platform, the central fraud detection platform at Anadolu Sigorta, was developed in-house using different technologies and methods, including KNIME Analytics Platform, Python, graph methods, and web service methodologies.

9 October 2024 in IEEE Access
Verified
Journal Article
Open access
FAFatih ABUT
Fatih ABUT
EAEzgi Akça
Ezgi Akça
MAMehmet Fatih Akay
Mehmet Fatih Akay
MIMustafa Irmak
Mustafa Irmak
MIMüzda Irmak
Müzda Irmak
YAYildirim Adigüzel
Yildirim Adigüzel
Resting Metabolic Rate (RMR) is a pivotal metric reflecting the caloric expenditure necessary for fundamental physiological functions during periods of rest, encompassing vital processes such as blood circulation, brain activity, respiration, fuel metabolism, and thermal regulation. The accurate estimation of RMR holds paramount importance in delineating an individual’s daily energy requirements and assessing their susceptibility to various health conditions, including cardiovascular diseases, hypertension, and diabetes. Additionally, RMR serves as a significant indicator of biological age, aiding in the formulation of personalized dietary plans by nutritionists to assist individuals in achieving their health and wellness objectives. This study aims to develop new models for predicting RMR using General Regression Neural Network (GRNN). Additionally, prediction models based on Multi-Layer Perceptron (MLP) will be created for comparison purposes. For model optimization, we developed an innovative web-based software using the DTREG predictive modeling software library. The implementation of this application was executed using Visual Studio and the C# programming language. To ensure the robustness of the prediction models, two distinct self-created datasets comprising data from female and male patients were employed. A total of 260 female patients and 150 male patients were included in the datasets. Separate GRNN-based and MLP-based models were developed for each dataset, leveraging various variables obtained from Tanita, a smart scale capable of comprehensive body scanning. These variables encompass personal information, body analysis, reference analysis, and segmental body composition analysis, providing detailed insights into impedance, fat percentage, fat mass, non-fat mass, muscle mass, and other metrics. The variables derived from Tanita measurements served as input variables for machine learning model creation, while RMR values obtained using FitMate™ served as output variables. Evaluation of these prediction models was based on Correlation Coefficient (R), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE) metrics to assess their performance, with the outcomes substantiating the efficacy of the proposed models in RMR prediction. The GRNN-based models performed better than the ones built by MLP, demonstrating its feasibility as a tool for predicting RMR accurately. The results have also been compared with up-to-date prediction equations from related literature, and our optimized models perform significantly better than those. The proposed best performing GRNN model achieved an R value of 0.85, RMSE of 134.91 kcal/day, and MAPE of 10.20%, significantly outperforming traditional predictive equations such as the Harris-Benedict and Mifflin-St Jeor equations, which showed RMSE values above 400 kcal/day and MAPE values exceeding 12.5%.