Sentiment Analysis (SA) concentrates on extracting and categorizing opinions from unstructured text, like product reviews or social media posts. It finds extensive application in brand evaluation, political campaigning, market analysis, and gathering customer feedback. A key method for SA involves supervised machine learning (SML), which relies on algorithms trained on labeled datasets to learn and identify patterns for classification. Among these tools, the “sentiment” package consistently delivers strong performance across various datasets. Although Python libraries generally offer faster processing speeds, it is important to recognize that while these packages effectively classify sentences as positive or negative, they often struggle with accurately capturing sentiment intensity. Our research indicates a common issue of overfitting, where these tools perform exceptionally well on known datasets but have difficulty with new, unseen data.