Detecting Salient features and Summarizing Health Review using Latent Dirichlet Analysis

Review means that “To examine something carefully especially before making decision or judgments”. Health consumers especially for health service providers author health reviews. Since the number of reviews are enormous, hence there is need to summarize these reviews. In this paper, we propose a simple approach to select the interesting topics of health consumers discuss when reviewing their health providers online. Our approach does not rely on any manual tagging of the information, and operates on the text of online reviews. We analyze a large set of reviews and find out the topics discussed when reviewing providers with different specialties. The health-rating information is based on the sentiment-classification result. The condensed descriptions of health reviews are generated from the feature-based summarization. We propose a novel approach based on Latent Dirichlet Analysis (LDA) to identify health features. Furthermore, we find a way to reduce the size of summary based on the health features obtained from LDA.