On exploring text mining approaches to sentiment analysis based on the combination of word-based and ontology-based approaches

Abstract

Currently, sentiment analysis plays an important role in business. Entrepreneurs try to understand customer needs for products and services. If they know about the needs, they can create the marketing plans or strategy plans in their business that help improve products and services. Therefore, this study explores two novel approaches to improve the classification accuracy of sentiment analysis data using a combination of a word-based approach (TF-IDF or CSDF) and an ontology-based approach (ontoSen) to provide two new methods, called ontoTF IDF and ontoCSDF. The experimental results show that CSDF method had the best classification accuracy among all the methods in this study: ontoCSDF did not improve further the classification accuracy of sentiment analysis data. Furthermore, ontoTFIDF method improved the classification by IBk algorithm significantly (p<0.05).