Main Article Content
Nowadays, there is a big trend in e-commerce websites where a large volume of valuable consumer sentiment data increases at an astonishing rate. Analyzing, summarizing, and explaining these big data (sentiments) quickly and efficiently is critical for data scientists. Many works exist in the public domain, such as hospitality, movies, hotel, electronic, and political reviews. Medical data analysis is a less explored area, and lots of research possibilities exist. Many people search for outsiders' opinions or views on different healthcare web portals, blogs and social sites related to a healthcare product, method, and services but cannot decide either sentiment is positive or negative. In this paper, we focus on attribute-level healthcare sentiment analysis. Firstly, design a Novel E-focused crawler and then pre-processing gathered reviews to remove noise content/outlier and improve storage capacity, efficiency, and Accuracy of the proposed system's relevant document. Secondly, corpus, Part of speech, thesaurus, and SentiWordNet dictionary are used to find out the implicit, explicit attributes and the polarity of sentiments.
At last, unsupervised clustering as Enhance K-Means (n-gram)-a machine learning approach is proposed. This proposed methodology can apply to several healthcare products and domains. The results obtained show our novel proposed approach outperformed all the existing methods in terms of Accuracy of 92%precision of 90 %, recall of 96%, F-Measure of 93%, G-Mean of 91%.