Cybersecurity Master's Student at Saarland University
This project explores analyzing speech data from COVID-19-related YouTube vlogs using Automatic Speech Recognition (ASR). The system matches symptoms from speech with known data to detect potential COVID-19 infections early. This approach aims to support pandemic response and public health surveillance.
Analyzed 22,524 English tweets tagged with #WorldCup2022 using VADER sentiment and word cloud analysis. While sports events triggered positive sentiments, the study also found strong negative sentiments regarding human rights concerns in the host nation.
Used NLP techniques to classify tweet sentiments about the Russia-Ukraine conflict. Applied TF-IDF for keyword analysis and Latent Dirichlet Allocation (LDA) for topic modeling, uncovering themes and emotional tones from public discourse.
Explored secure computations on encrypted data using PYFHEL and Concrete-Numpy. Applied encryption to small datasets and executed NumPy operations without decrypting. Demonstrated the practicality and challenges of privacy-preserving machine learning.
Developed two tools for MUET students: