RESEARCH EXPERIENCE

Master's Dissertation

TweetGuard: Combining Transformer and Bi-LSTM Architectures for Fake News Detection in Large-Scale Tweets

Developed "TweetGuard," a novel classification model for detecting fake news in tweets, integrating transformer and LSTM architectures. Utilized the "TruthSeeker" dataset for robust training, enhanced model performance using advanced text cleaning and BERTweet tokenization. Validated robustness across multiple datasets, significantly outperforming traditional deep learning classifiers and advancing fake news detection technology.

Undergraduate Thesis

Feature Extraction & Classification of Electrooculography (EOG) Signal Using RNN for Human Computer Interface (HCI)

This work analyzed the performance of RNN-based deep learning algorithms using feature extraction on a EOG dataset to identify the four directional eye movements. This work used MATLAB for signal denoising and baseline drift mitigation along with feature extraction as well as Python 3.7 programming language with Scikit-Learn under Google Colab distribution for classification.

Google Sites

Report abuse