Encrypting sentiments: a study on integrating encryption module with NLP pipeline to analyze emotions while ensuring security

Abstract

This research delves into the critical role of encryption in Natural Language Processing (NLP), emphasizing its significance as an emergency text platform, akin to a text-based emergency broadcast service. The study analyzes sentiments such as sadness, neutrality, worry, love, surprise, etc., utilizing a standard NLP pipeline for sentiment analysis. Additionally, it compares the results with an enhanced approach that incorporates an encryption module, aiming to quantify potential data loss in the latter scenario and highlighting the trade-offs between data protection and sentiment analysis accuracy in NLP. Addressing the prevalent absence of security components in existing NLP pipelines, this research introduces encryption to enhance security. This academic pursuit sheds light on the nuanced relationship between data protection and sentiment analysis accuracy in the context of NLP, providing valuable insights to guide the refinement of more resilient emergency text-based services while creating a cross-section between the NLP pipeline and Encryption Module.