A NOVEL SIMILARITY APPROACH FOR ONLINE SENTIMENT TEXT

Computer Engineering

The content and context of social network websites become crucial to know what the people are interested in and what kinds of information are spread among them depending on all commissions, comments, and actions need to analyze. Consequently, it becomes important that the brands listen carefully to what is said about their online business. Additionally, it demands more challenges to know whether the conversation leads to positive or negatives so that the impact of social network opinions can be measured to apply back in the real word problems. This paper finds the similar groups of social network activities, especially comments and posts of the users who shares about the same context depending on a specific topic. For this purpose, this paper introduces how to deal with finding the similarity between the contextual text of the users in semantic ways by filling the gap of syntactic measures in text similarity. Regarding the datasets, Twitter dataset, which is a popular dataset for sentiment analysis is used Respecting to the performance results, the proposed system achieves promising results with higher accuracy rate but lower error rate for both datasets available from online.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.