Sentiment analysis using healthcare data

Abstract

Listen

Translate article

Healthcare sentiment analysis focuses on diagnosing healthcare-related issues that people have discovered. To create rules and changes that could directly address patients' issues by considering their input. Machine learning techniques analyze millions of review documents and conclude them towards an efficient and accurate decision. This study trained the health data using different machine-learning algorithms: Multinomial Naïve Bayes, Random Forest, Decision Trees, K-nearest neighbor, and Support Vector Machine and compared the accuracy using different evaluation metrics. The study used a drug review dataset from the UCI machine-learning repository. It was separated so that 70% of the data was used as a training dataset for the ML models. The remaining 30% of the data forms the test dataset used to evaluate the trained ML models. Based on the evaluation metrics, the random forest has the highest accuracy (89.4%) and R squared (0.501), and the lowest MSE (10.5) and RSME (0.324). The study concluded that the random forest classifier is the optimal model for predicting healthcare data while KNN has the lowest accuracy. It is recommended government health ministry and healthcare facilities use online health data to create policies that will directly address these public health issues, allowing patients to directly address their concerns to higher authorities without having to go through arduous procedures