Undergraduate Research Assistant
As an Undergraduate Research Assistant, I accomplished a 25% increase in model efficiency. I achieved 88% accuracy in parsing keywords from social media posts by designing and training an NLP model in Python. I improved the model’s performance by training it on over 38,000 data entries using a snowball iterative algorithm to effectively interpret human language in social media. Additionally, I extracted key data from a secondary dataset of 600 entries to support accurate post-identification. I collaborated with a team of two researchers and a supervisor to integrate the model into an automated social media bot aimed at informing users about food safety practices. Our research findings were presented at NCUR 2023 and Edison STEM-NET Research 2023, showcasing the impact of our work on food safety and advancements in NLP.