Comparative study of service-based sentiment analysis of social networking sites fanatical contents

Abstract

The proliferation of mobile web services (MWS) for sentiment analysis makes it hard to identify the best MWS for sentiment analysis of social networking sites’ fanatical contents. This paper carries out a comparative study of service-based sentiment analysis of social networking sites’ fanatical contents. This is achieved by cleaning, transformation, and reduction of fanatical contents from the publicly available social media dataset, and multiple MWS are selected for comparison using the application programming interface (API) key of the MWS. To evaluate the service-based sentiment analysis, standard measures such as accuracy, precision, recall, and f-measures of sentiment result for each MWS are used. The result shows that Dandelion SA performs better in terms of accuracy (72.5%) and recall (76.9%), while Wingify SA performs better in terms of precision (88.6%) and f-measure (75.5%), though AlchemyAPI offers the most crucial elements in analyzing sentiments such as emotion, relevance score, and sentiment type. The outcomes of this paper will benefit the sentiment analysis service developers, sentiment analysis service requesters as well as other researchers in the social media fanatical content domain.