Computational Modelling of World Leaders’ Covid -19 Opinions: A Sentiment Analysis Approach

Sentiment analysis uses the processing of Natural language Processing (NLP) together with analysis of text and linguistic patterns to help in identifying the state of information which has been mined from reliable sources. The polarization of a textual pattern, whether a linguistic pattern is positive, neutral or negative can be evaluated with the help of computational modelling of the pattern. Performing computational modelling on sentiment analysis can bring out useful patterns which can be employed for monitoring purposes in social media. Sentiment analysis also finds its vast applications in the management and social aspects of sciences due to its varied importance to commercial applications and society. The growing prominence of sentiment analysis matches with the exponential progress of social media such as Facebook, Twitter, and other social networks. This research paper recommends a standalone platform which allows data handlers to execute sentiment analysis which employs Twitter API to collect information about COVID- 19 related tweets of world leaders. The proposed model is used to analyse whether the sentiment in a tweet is positive, neutral or negative. The pre-processed data passes through the sentiment generation phase where sentiment classes are demarcated for polarity. In this research paper, tweets of various world leaders are labelled with sentiments, and the overall sentiment of their tweets are analysed.

I. Introduction

Sentiment analysis is the method of analysing the sentiments, opinions, appraisals and attitudes of people towards entities like products, individuals, events, social issues or/and their attributes. As it can represent a large problem space, it can be communicated through a portion of text, to achieve the writer’s/speaker’s polarity towards an idea by means of the model of computational linguistics [1]. The portion of text could include an issue or a topic or a product on which the user can give his/her own opinions [2]. This can be used to analyse requirements from a customer/client, deduce his needs and acquire judicious feedback. Sentiment analysis is widely used in recommender systems to predict preference of one object over another. It is also considered as research technique to give a systematic and complete depiction of data, organized into a hierarchical organization, ordered in terms of comparative importance with existing alternates.