An Enhanced Exploration of Sentimental Analysis in Health Care

Abstract

The medical dataset replicates the patient's crucial information, such as important details regarding the patient's health. It includes disease diagnoses, interventions, and descriptions of the examined results. Also, detecting the mindset of an acute disease-affected patient is a primary challenging task. Though sentiment analysis plays a role in seeing their perspective, the significant broad medical application does not yet meet the analysis of patient mindset. So here we identified major shortcoming exists while studies the diversified disease-affected people mindset. Hence, we introduce a practical framework to analyse patients' perspectives using a socio-medical dataset that contains various reviews and feedback of critical diseases-affected people—initially, we applied a pre-processing technique, including Lowercase Conversion, removing special characters, removing stop words, Number to word conversion, Stemming, and lemmatization over dataset. Next, N-gram tokenization methodology is used to extract the valuable features followed by assigning polarity score to each sentiment we extract and calculate the overall polarity of the context. Finally, a probabilistic LDA model was employed to combine the review. Furthermore, various machine learning classifiers are explored to evaluate the performance of the proposed framework.