Enhancing Optimized Personalized Therapy in Clinical Decision Support System using Natural Language Processing

Abstract

Sentiment analysis is the process of identifying and categorising the opinions expressed by human utterances through computational techniques using natural language processing. The present work focuses on a case study to develop a clinical decision support system for personalized therapy process using aspect-based sentiment analysis. The process is carried out on a drug review data in order to determine whether the patient’s behaviour towards a medicine, product, treatment etc is positive, negative or neutral using NLP techniques. The polarities obtained are compared for further analysis of the patient reviews for the better clinical decision system. Machine learning methods are also used for classification of the drug review data to compare the sentiment scores. The prominent statistical sklearn models used are support vector machines (SVM), Random Forest Classification, LinearSVC, MultinomialNB. SVM algorithm is found to perform better compared to other in terms of accuracy.

Keywords

• Polarity
• Tokenizer
• Sentiment score
• Sentiment label
• Classification
• Natural language processing