Analysing Domain-Specific Sentiments in Healthcare: SentiWordNet Adjusted Vader Sentiment Analysis (SAVSA)

Abstract

Sentiment analysis, the process of understanding people's emotions and opinions in the digital age, holds significant importance for both content creators and consumers in the era of abundant online content. However, accurately gauging sentiment can be challenging, particularly when individuals employ specialized language and online shorthand. This research work presents a novel approach to sentiment analysis known as SentiWordNet Adjusted Vader Sentiment Analysis (SAVSA). SAVSA leverages various tools, including SentiWordNet and Vader, in conjunction with the domain expertise of SenticNet7, to distinguish the emotional tones conveyed by text. Unlike traditional sentiment analysis methods that predominantly focus on general sentiment, SAVSA is specifically tailored to excel in the medical domain, addressing real-world challenges that conventional tools like Vader and SentiWordNet often struggle to tackle effectively. SAVSA emerges as a robust and adaptable method for comprehending emotional tones within textual content. Its applications span diverse domains, from assessing public sentiment regarding medical topics to analyzing consumer sentiments related to online shopping experiences. Furthermore, SAVSA proves valuable for examining the emotions expressed on social media platforms like Twitter, especially during significant events such as the rollout of the COVID-19 vaccine. To evaluate the efficacy of SAVSA, a comprehensive comparative analysis was conducted, comparing its performance with other sentiment analysis tools. This research endeavors to introduce an innovative approach that excels in deciphering how people express their emotions online, particularly in specialized domains. The need for such specialized sentiment analysis arises from the inherent complexity of discussing specialized topics and SAVSA holds the promise of simplifying this process.