3 research outputs found
Big Data in Cybersecurity: Enhancing Threat Detection with AI and ML
No full textThe growing sophistication and number of cyber threats have made it imperative to incorporate big data analytics, artificial intelligence (AI), and machine learning (ML) in cybersecurity. This study investigates AI-based models for improved threat detection, with emphasis on Random Forest, Support Vector Machines (SVM), Deep Learning, and K-Means Clustering. The research employs a dataset of 500,000 cybersecurity incidents, examining attack patterns, anomaly detection, and fraud prevention systems. Experimental outcomes prove that the Deep Learning model exhibited maximum accuracy at 96.8%, surpassing SVM at 92.3% and Random Forest at 94.1% for the detection of ransomware and intrusion attempts. K-Means Clustering also successfully classified malicious behavior at a detection level of 89.5%. Outcome shows that AI-based methods substantially improve real-time cyber threat mitigation over conventional approaches. In addition, the use of blockchain and big data analytics enhances financial transaction fraud detection by 35% less false positives. AI and ML, the research concludes, provide better accuracy, flexibility, and velocity in cybersecurity uses. Computational cost and adversarial attacks are the challenges that need to be optimized. More interpretable and scalable AI models need to be developed in future studies to improve global cybersecurity resilience
A novel switched-capacitor multilevel inverter for efficient voltage level generation
No full textThis paper presents a novel single direct current (DC) source with switched-capacitor multilevel inverter (MLI) architecture capable of achieving seven output voltage levels using only eight switches, one diode, and two capacitors. The proposed topology (P) is compared with recent MLI configurations to assess its efficiency and performance. MATLAB/Simulink tools are utilized for simulation studies, and experimental validation is conducted to corroborate the theoretical findings. The investigation explores the impact of modulation index and switching frequency variations on the P output characteristics. Results indicate that the proposed MLI topology offers significant advantages in terms of component count reduction and simplicity while maintaining competitive performance compared to state-of-the-art alternatives. Additionally, the study provides insights into the influence of modulation index and switching frequency changes on the output voltage waveform, highlighting the adaptability and robustness of the P under varying operating conditions. This research contributes to the advancement of MLI designs by offering a streamlined and efficient solution suitable for various power electronic applications, including renewable energy systems and motor drives, where minimizing component count and complexity are crucial design considerations
