Long Short Term Memory-based Sentiment Analysis of Healthcare Reviews

Reviewing the posts and feedbacks of different users on various services or products has become a major concern these days due to the rapid increase in those textual data over the web. The manual processing of such huge data is time consuming, therefore found to be less effective in quick decision making by different service providers or customers. Automatic analysis of those contents can provide a platform for the service providers to assist in decision-making and formulating different effective plans for improvising their services as and when needed. This may also help the customers to reach to a decision quickly. The area of healthcare services is one of the most demanding areas where very often people search for web information about doctors, hospitals, medicines, and other services. Automatic sentiment analysis of available customer feedbacks and reviews may help healthcare service providers to make service improvements against their competitors and may help the customers to take decisions quickly. In this paper, we present a model for text sentiment analysis of the healthcare reviews using Long short-term memory (LSTM). The presented model is tested on various healthcare reviews collected from different internet sources and the performance of the model is evaluated with other models. The obtained results indicate the potentiality of the model in findina the aunrouriate sentiment.