SÇSemih Çakır

Semih Çakır

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Semih Çakır has a total of 3 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Enhanced Prediction of Gain and Bandwidth for Metamaterial Antenna Using Machine Learning and Metaheuristic Optimization

    by Nesibe Yalçın, Sibel Ünaldı, Semih Çakır

    The ongoing demand for improved performance in modern communication systems has driven recent developments in electromagnetic component design. Integrating metamaterial unit cell structures into antenna designs can be a solution for improving electromagnetic radiation properties of antennas. Also, in the antenna design process, both the achievement of multiple performances (such as size, gain, bandwidth, and return loss) and time efficiency have become increasingly important, and this is accomplished through advanced techniques like Machine Learning (ML) integrated with optimization algorithms. The purpose of this research is to develop an ML model for estimating the performance of a metamaterial antenna (gain and bandwidth). The Particle Swarm Optimization heuristic algorithm is used to optimize the hyperparameters of the ML models (Random Forest and Support Vector Machine). The ML models were evaluated using a comparison table in terms of R2, MAE, MSE, and RMSE. The PSO-SVM hybrid model showed better predictive performance than PSO-RF, with MSEs of 0.0003 for gain and 0.0020 for bandwidth. The optimized ML models can be used to predict antenna gain and bandwidth more accurately.

  • Çeşitli Tekrarlayan Sinir Ağları Kullanarak Siber Saldırı Tespiti

    by Nesibe Yalçın, Semih Çakır, Sibel Ünaldı

    Siber saldırıların erken tespiti ve tanımlanması; saldırıların etkisinin doğru bir şekilde değerlendirilmesi, bunlara karşı hızlı ve etkili önlemlerin alınması, veri ve sistemlerin korunması, iş sürekliliğin sağlanması, kurumsal itibarın korunması, yasal ve düzenleyici standartlara uyumun sağlanması açısından hayati önem taşımaktadır. Bu çalışmada, bir siber saldırı tespit sistemi önerilmiştir. Saldırı tespiti için çeşitli Tekrarlayan Sinir Ağı (Recurrent Neural Network, RNN) derin öğrenme yöntemlerinin yanı sıra K En Yakın Komşu (K Nearest Neighbour, KNN), Karar Ağacı ve Rastgele Orman makine öğrenme algoritmaları uygulanmıştır. Sistemin saygınlığı, KDD’99 veri setinin %10’u üzerinden değerlendirilmiş ve tartışılmıştır. Öğrenme modellerinin başarımını karşılaştırmak için çeşitli değerlendirme metrikleri kullanılmıştır. Aynı veri setini kullanan çalışmalarla karşılaştırıldığında önerilen Çift Yönlü Uzun Kısa Süreli Bellek, daha yüksek başarıma sahip RNN modeli olarak öne çıkmaktadır. Ayrıca KNN de yüksek bir test doğruluğu (%99.92) ve duyarlılık (%99.94) sunmuştur. Önerilen modeller, siber saldırıların erken aşamada tespit edilmesini kolaylaştırabilir.


  • Open Access

    Attack Detection Using Artificial Intelligence Methods for SCADA Security

    by Nesibe Yalçın, Semih Çakır, Sibel Ünaldı

    Technological developments and transformations have rapidly risen since the Fourth Industrial Revolution. The prevalence of industrial devices interconnected over the wireless sensor networks and the provision of a sustainable data flow reveal the importance of the Industrial Internet of Things (IIoT). In the manufacturing industry, supervisory control and data acquisition (SCADA) systems are used to control IIoT for critical infrastructure. A cyberattack on the network-based communication structure embedded into the architecture of industrial equipment can significantly disrupt/sabotage product manufacturing and other industrial operations. The digitization of industrial control systems can expose the systems to malicious actors and therefore requires additional security solutions, such as intrusion detection systems (IDSs). Increasing sophistication of cyberattacks, industrial companies need to adopt innovative solutions like artificial intelligence (AI)-based attack detection to protect their valuable assets. In addition, AI-based approaches are more effective as they analyze network traffic, identify threats, and adapt to new attack techniques. This study aims to develop an AI-based IDS with high accuracy for SCADA security. In the study, cyberattacks that may occur against SCADA systems are examined. AI methods (including K-nearest neighbor, quadratic discriminant analysis, adaptive boosting, gradient boosting, and random forest) in different categories are used and AI models with various parameters are built. To improve the detection performance of the models, comprehensive experiments are carried out on two different SCADA data sets. As a result of experiments, the test accuracy rates exceeding 96.82% are achieved by all models: on the WUSTL-IIOT-2021 data set, the XGB model has outperformed with an accuracy of 99.99%.


  • Open Access

    RPL Attack Detection and Prevention in the Internet of Things Networks Using a GRU Based Deep Learning

    by Semih Cakir, Sinan Toklu, Nesibe Yalcin

    Cyberattacks targeting Internet of Things (IoT), have increased significantly, over the past decade, with the spread of internet-connected smart devices and applications. Routing Protocol for Low-Power and Lossy Network (RPL) enables messages to be routed between nodes for the Wireless Sensor Network in the network layer. RPL protocol, which is sensitive and difficult to protect, is exposed to various attacks. These attacks negatively affect data transmission and cause great destruction to the topology by consuming the resources. Hello Flooding (HF) attacks against RPL cause consumption of constrained resources (memory, processing and energy) in nodes. Therefore, in this study, a Gated Recurrent Unit network model based deep learning has been proposed to predict and prevent HF attacks on RPL protocol in IoT networks. The proposed model has been compared with Support Vector Machine and Logistic Regression methods, and different power states and total energy consumptions of the nodes have been taken into consideration and experimented with. The results confirm the promised and expected performance from the model in terms of source efficiency and IoT security. In addition, attack detection has been carried out with a much lower error rate than literature studies for HF attacks from RPL flood attacks.