A Framework for Evaluating Medical Blog and Camera Opinions Based on Opinion Mining and Sentiment Analysis

摘要

Opinion mining also called sentiment analysis is a process of finding users opinion about particular topic. The key challenge faced in opinion mining is that the natural language is highly unstructured in nature and interpretation of the meaning of a particular word, phrase or sentenceby a machine is cumbersome. But the usefulness of the sentiment analysis is increasing day by day as large source of user generated contents (in the form of blogs, comments, reviews, wikis) act as important source for web mining which can be used for product feedback analysis, and for decisionmaking to users. In this work, the efficiency of the feature extraction methods and classification algorithms for classifying cameras reviews were investigated. Opinions expressed on cameras are taken from Amazon website. TDF×IDF is utilized for the extraction of features from camerareviews. Features transformation is undertaken by using PCA and kernel PCA. Three classification algorithms Na?ve Bayes, K Nearest Neighbour and Classification and Regression Trees (CART) algorithms were used to investigate the quality of the extracted features. Experimental results demonstratethat features extracted using TDF×IDF with kernel PCA enhances the classification precision of the classifiers. Outcomes reveal that CART algorithm has higher classification accuracy than other classifiers.