Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/395
Browse
15 results
Search Results
Article Citation - Scopus: 7Network Anomaly Detection Using Deep Autoencoder and Parallel Artificial Bee Colony Algorithm-Trained Neural Network(PeerJ Inc., 2024-10-08) Dedeturk, Bilge Kagan; Bakir-Gungor, Burcu; Hacılar, Hilal; Gungor, Vehbi CagriArticle Citation - Scopus: 1Robust Controller Electromyogram Prosthetic Hand With Artificial Neural Network Control and Position(Indian Journal of Forensic Medicine and Toxicology ijfmt@hotmail.com, 2020) Ahmed, Saygin Siddiq; Ahmed, Aydin S.; Yilmaz, Bulent; Doǧru, NuranIn this study, we proposed and designed a new control method for an electromyographically (EMG) controlled prosthetic hand. The objective is to increase the control efficiency of the human–machine interface and afford greater control of the prosthetic hand. The process works as follows: EMG biomedical signals acquired from Myoware sensors positioned on the relevant muscles are sent to the robot that consist of hand, Arduino and MATLAB program, which computes and controls the hand position in free space along with hand grasping operations. The Myoware device acquires muscle signals and sends them to the Arduino. The Arduino analyzes the received signals, based on which it controls the motor movement. In this design, the muscle signals are read and saved in a MATLAB system file. After program processing on the industrial hand which is applied by MATLAB simulation, the corresponding movement is transferred to the hand, enabling movements, such as, hand opening and closing according to the signal stored in the MATLAB system. In this study, hand and fingerprints were designed using a three-dimensional printer by separate recording finger and thumb signals. The muscle signals were then analyzed in order to obtain peak signal points and convert them into data. These results indicate the effectiveness of the proposed method and demonstrate the superiority of the method for amputees because of the improved controllability and perceptibility afforded by the design. © 2020 Elsevier B.V., All rights reserved.Conference Object Citation - WoS: 7Citation - Scopus: 10PI-Controlled ANN-Based Energy Consumption Forecasting for Smart Grids(SciTePress, 2015) Gezer, Gülsüm; Tuna, Gürkan; Κogias, DImitrios G.; Gülez, Kayhan; Güngör, Vehbi Çağrı; Kogias, DimitrisAlthough Smart Grid (SG) transformation brings many advantages to electric utilities, the longstanding challenge for all them is to supply electricity at the lowest cost. In addition, currently, the electric utilities must comply with new expectations for their operations, and address new challenges such as energy efficiency regulations and guidelines, possibility of economic recessions, volatility of fuel prices, new user profiles and demands of regulators. In order to meet all these emerging economic and regulatory realities, the electric utilities operating SGs must be able to determine and meet load, implement new technologies that can effect energy sales and interact with their customers for their purchases of electricity. In this respect, load forecasting which has traditionally been done mostly at city or country level can address such issues vital to the electric utilities. In this paper, an artificial neural network based energy consumption forecasting system is proposed and the efficiency of the proposed system is shown with the results of a set of simulation studies. The proposed system can provide valuable inputs to smart grid applications. © 2022 Elsevier B.V., All rights reserved.Article Citation - WoS: 55Citation - Scopus: 65Neuro-Fuzzy Model Predictive Energy Management for Grid Connected Microgrids(MDPI, 2020-05-28) Ulutas, Ahsen; Altas, Ismail Hakki; Onen, Ahmet; Ustun, Taha SelimWith constant population growth and the rise in technology use, the demand for electrical energy has increased significantly. Increasing fossil-fuel-based electricity generation has serious impacts on environment. As a result, interest in renewable resources has risen, as they are environmentally friendly and may prove to be economical in the long run. However, the intermittent character of renewable energy sources is a major disadvantage. It is important to integrate them with the rest of the grid so that their benefits can be reaped while their negative impacts can be mitigated. In this article, an energy management algorithm is recommended for a grid-connected microgrid consisting of loads, a photovoltaic (PV) system and a battery for efficient use of energy. A model predictive control-inspired approach for energy management is developed using the PV power and consumption estimation obtained from daylight solar irradiation and temperature estimation of the same area. An energy management algorithm, which is based on a neuro-fuzzy inference system, is designed by determining the possible operating states of the system. The proposed system is compared with a rule-based control strategy. Results show that the developed control algorithm ensures that microgrid is supplied with reliable energy while the renewable energy use is maximized.Article Citation - WoS: 4Citation - Scopus: 7Network Anomaly Detection Using Deep Autoencoder and Parallel Artificial Bee Colony Algorithm-Trained Neural Network(PeerJ Inc, 2024-10-08) Hacilar, Hilal; Dedeturk, Bilge Kagan; Bakir-Gungor, Burcu; Gungor, Vehbi CagriCyberattacks are increasingly becoming more complex, which makes intrusion detection extremely difficult. Several intrusion detection approaches have been developed in the literature and utilized to tackle computer security intrusions. Implementing machine learning and deep learning models for network intrusion detection has been a topic of active research in cybersecurity. In this study, artificial neural networks (ANNs), a type of machine learning algorithm, are employed to determine optimal network weight sets during the training phase. Conventional training algorithms, such as back- propagation, may encounter challenges in optimization due to being entrapped within local minima during the iterative optimization process; global search strategies can be slow at locating global minima, and they may suffer from a low detection rate. In the ANN training, the Artificial Bee Colony (ABC) algorithm enables the avoidance of local minimum solutions by conducting a high-performance search in the solution space but it needs some modifications. To address these challenges, this work suggests a Deep Autoencoder (DAE)-based, vectorized, and parallelized ABC algorithm for training feed-forward artificial neural networks, which is tested on the UNSW-NB15 and NF-UNSW-NB15-v2 datasets. Our experimental results demonstrate that the proposed DAE-based parallel ABC-ANN outperforms existing metaheuristics, showing notable improvements in network intrusion detection. The experimental results reveal a notable improvement in network intrusion detection through this proposed approach, exhibiting an increase in detection rate (DR) by 0.76 to 0.81 and a reduction in false alarm rate (FAR) by 0.016 to 0.005 compared to the ANN-BP algorithm on the UNSWNB15 dataset. Furthermore, there is a reduction in FAR by 0.006 to 0.0003 compared to the ANN-BP algorithm on the NF-UNSW-NB15-v2 dataset. These findings underscore the effectiveness of our proposed approach in enhancing network security against network intrusions.Conference Object Citation - Scopus: 1Improving Salary Offer Processes With Classification Based Machine Learning Models(Institute of Electrical and Electronics Engineers Inc., 2024-09-21) Kaya, Rukiye; Saatci, Mehtap; Bakal, Gokhan; Bakal, Mehmet GokhanIn job applications, salary is major motivational factor for employees and making accurate salary prediction is crucial for both employers and employees. Utilizing advanced technologies can significantly enhance the accuracy and efficiency of salary prediction process. In this study, we explore Machine Learning (ML) methods to enhance salary prediction process. We evaluated seven classification models for predicting salary categories, with the Artificial Neural Network (ANN) model achieving the highest accuracy at 58.2% on the test dataset, followed by the K-Nearest Neighbors (KNN) model with an accuracy of 56.8%. Additionally, we employed ensemble models to further enhance prediction accuracy. Among these, the Majority Voting Classifier using Hard Voting achieved the highest accuracy at 59.3%, demonstrating the potential of ensemble techniques in refining salary predictions. The developed salary prediction tool estimates the most appropriate salary category for each candidate and help mitigate potential biases in manual salary assessments, hence enables a more objective and consistent compensation system. ∗CRITICAL: Do Not Use Symbols, Special Characters, or Math in Paper Title or Abstract, and do not cite other papers in the abstract. © 2024 Elsevier B.V., All rights reserved.Article Citation - WoS: 34Citation - Scopus: 37Green Building Envelope Designs in Different Climate and Seismic Zones: Multi-Objective ANN-Based Genetic Algorithm(Elsevier, 2022-10) Himmetoglu, Salih; Delice, Yilmaz; Aydogan, Emel Kizilkaya; Uzal, Burak; Kızılkaya Aydoğan, EmelIn recent years, the major component of green building designs adopted by governments in order to reduce CO2 emissions as well as energy consumption is the green building envelope. The green envelope has the most important share in terms of thermal energy consumption, environment, and indoor comfort criteria. Determining the most suitable building envelope combination in the building life cycle is an important problem for designers. This study presents a new multi-objective approach that determines the most suitable green envelope designs for the buildings in different climate and earthquake zones, taking into account CO2 emissions, heating/cooling energy consumption, and material cost in terms of life cycle cost analysis. To this end, EnergyPlus building performance simulation program, artificial neural network (ANN), and genetic algorithm are used together. After the heating and cooling energy consumption, CO2 emissions, and material cost values are obtained for a certain number of the envelope alternatives with the EnergyPlus, ANN models that learn the working mechanism of EnergyPlus are trained according to these values. An ANN-based genetic algorithm procedure is developed to search the whole envelope alternative space by using the trained ANN models with EnergyPlus. The proposed approach allows searching in a very short time the whole alternative space, which is almost impossible to scan with EnergyPlus by reducing the time spent and the number of alternatives required for the design and simulation processes of the green building envelope. The proposed approach is performed for a design-stage city hospital structure in Turkey. Window type, the internal/external plaster, wall, and insulation materials along with the thicknesses of these materials, which consist of 46 different variables, are determined as envelope attributes for four different climate and seismic zones. The green building envelope designs obtained with the proposed approach are entered into EnergyPlus and the consistency of the results is compared. ANN models with an average accuracy of over 97% are developed. Without the CO2 emission cost in the life cycle cost, the mean absolute percent error (MAPE) values for each region are 0.67%, 0.6%, 0.58%, and 1.78%, respectively. With the CO2 emission cost in life cycle cost, the MAPE values for each region are 0.96%, 0.88%, 0.86%, and 0.43%, respectively. According to the obtained results, there is a consistency of over 99% between EnergyPlus and the proposed approach.Article Citation - Scopus: 8Building a Challenging Medical Dataset for Comparative Evaluation of Classifier Capabilities(Elsevier Ltd, 2024-08) Bozkurt, Berat; Coskun, Kerem; Bakal, GokhanSince the 2000s, digitalization has been a crucial transformation in our lives. Nevertheless, digitalization brings a bulk of unstructured textual data to be processed, including articles, clinical records, web pages, and shared social media posts. As a critical analysis, the classification task classifies the given textual entities into correct categories. Categorizing documents from different domains is straightforward since the instances are unlikely to contain similar contexts. However, document classification in a single domain is more complicated due to sharing the same context. Thus, we aim to classify medical articles about four common cancer types (Leukemia, Non-Hodgkin Lymphoma, Bladder Cancer, and Thyroid Cancer) by constructing machine learning and deep learning models. We used 383,914 medical articles about four common cancer types collected by the PubMed API. To build classification models, we split the dataset into 70% as training, 20% as testing, and 10% as validation. We built widely used machine-learning (Logistic Regression, XGBoost, CatBoost, and Random Forest Classifiers) and modern deep-learning (convolutional neural networks - CNN, long short-term memory - LSTM, and gated recurrent unit - GRU) models. We computed the average classification performances (precision, recall, F-score) to evaluate the models over ten distinct dataset splits. The best-performing deep learning model(s) yielded a superior F1 score of 98%. However, traditional machine learning models also achieved reasonably high F1 scores, 95% for the worst-performing case. Ultimately, we constructed multiple models to classify articles, which compose a hard-to-classify dataset in the medical domain. © 2024 Elsevier B.V., All rights reserved.Article Citation - WoS: 4Citation - Scopus: 4Beneficiation of Low-Grade Iron Ore Using a Dry-Roll Magnetic Separator and Its Modeling via Artificial Neural Network(Springer, 2025-02-24) Fariss, Abdourahman Hassan Brahim; Ibrahim, Ahmedaljaali Ibrahim Idrees; Ozdemir, Ali Can; Top, Soner; Kursunoglu, Sait; Altiner, MahmutThe beneficiation of low-grade iron ore (39.5% Fe-(T) grade) using a dry-roll magnetic separator was investigated. The ore was characterized using Mineral Liberation Analysis (MLA). It was determined that the ore was composed of iron oxide (goethite and hematite), quartz, chlorite, muscovite, plagioclase, and other minerals. The effect of particle size (PS, - 1 + 0.500 mm, - 0.500 + 0.300 mm, and - 0.300 + 0.125 mm), splitter position (SP, 43 degrees and 58 degrees), cleaning stage (CS, 1 and 2), conveyor speed (CoS, 3, 5, and 7 Hz), magnetic field strength (MFS, 0.2 T and 0.4 T) on the recovery of the magnetic product was investigated. Experimental results show that the product (- 1 + 0.500 mm) with the Fe-(T) grade of 67.67% can be obtained, but its recovery was not at an acceptable value (< 30%). Furthermore, the Fe-(T) grade of the product (- 0.500 + 0.300 and - 0.300 + 0.125 mm) could not reach satisfactory levels<bold>.</bold> The artificial neural network (ANN) method was conducted on the results of experimental studies. Three different training algorithms were employed for modeling, and their performance was assessed using statistical evaluation criteria. The results demonstrate that Bayesian Regularization (BR) algorithm exhibited better performance compared to others in predicting both Fe(T) grade and recovery rate during the testing phase. These findings support the notion that ANN algorithms can be a powerful modeling and prediction tool in the field of mineral processing.Article Citation - WoS: 6Citation - Scopus: 7Assessment of Los Angeles Abrasion Value (LAAV) and Magnesium Sulphate Soundness (MWL) of Rock Aggregates Using Gene Expression Programming and Artificial Neural Networks(Polska Akad Nauk, Polish Acad Sciences, 2023-07-24) Koken, EkinIt has been acknowledged that two important rock aggregate properties are the Los Angeles abrasion value (LAAV) and magnesium sulphate soundness (Mwl). However, the determination of these properties is relatively challenging due to special sampling requirements and tedious testing procedures. In this stu-dy, detailed laboratory studies were carried out to predict the LAAV and Mwl for 25 different rock types located in NW Turkey. For this purpose, mineralogical, physical, mechanical, and aggregate properties were determined for each rock type. Strong predictive models were established based on gene expression programming (GEP) and artificial neural network (ANN) methodologies. The performance of the proposed models was evaluated using several statistical indicators, and the statistical analysis results demonstra-ted that the ANN-based proposed models with the correlation of determination (R2) value greater than 0.98 outperformed the other predictive models established in this study. Hence, the ANN-based predictive models can reliably be used to predict the LAAV and Mwl for the investigated rock types. In addition, the suitability of the investigated rock types for use in bituminous paving mixtures was also evaluated based on the ASTM D692/D692M standard. Accordingly, most of the investigated rock types can be used in bituminous paving mixtures. In conclusion, it can be claimed that the proposed predictive models with their explicit mathematical formulations are believed to save time and provide practical knowledge for evaluating the suitability of the rock aggregates in pavement engineering design studies in NW Turkey.