Abstract

Within the field of natural language processing, sentiment analysis is one form of data mining used to make inferences about the emotional tenor of a speakers words. Computational linguistics is employed to examine the text in order to deduce and assess ones mental knowledge of the Web, social media, and associated references. One of the numerous advantages of sentiment analysis is that it can help improve the quality of healthcare by making use of medical data to produce the most positive outcome possible. Natural language processing challenges can change how sentiment analysis looks and works in a variety of contexts. Some of the challenges are specific to the data type, while others are universal to any method of text analysis. The primary objective of this study was to evaluate how challenging it is to analyse sentiment in the healthcare sector. Given the aforementioned complexities, the objective was to look into whether or not the currently available SA tools are adequate for handling any healthcare-related issue. With such motivation, in this paper, we develop an unsupervised ensemble machine learning (ML) algorithm that includes K-means clustering; Principle Component Analysis; Independent Component Analysis and k-nearest neighbors. The unsupervised ensemble ML model is assessed via voting meta-classifier over various cancer datasets. The simulation is conducted to test the efficacy of the model in terms of accuracy, precision, recall and f-measure over various datasets. The results of simulation against the cancer datasets show that the proposed method achieves higher rate of ensemble accuracy than the other existing ensemble models.