Abstract

Vast amounts of new types of data are constantly being created as a result of dynamic digitization in all areas of our lives. One of the most important and valuable categories for business is data from social networks such as Facebook. Feedback resulting from the sharing of thoughts and emotions, expressed in comments on various products and services, is becoming the key factor on which modern business is based. This feedback is called sentiment on social media. The sentiment analysis method is used for the mining of polarity views, feelings and emotions from text data with the help of natural language processing (NLP). The fundamental task in this technique is to classify the polarity of opinions into classes such as positive, negative and neutral. This chapter introduces a novel approach to the quantitative measurement of sentiment by means of a created indicator called the Net Sentiment Rate (NSR). The proposed NSR expresses the net sentiment extracted from text data and provides values on a scale from -1 to +1. The Net Sentiment Rate implementation has been verified on large datasets crawled from Facebook in the period from 1 October 2014 to 31 December 2018, for three international companies from the same sector. The created Net Sentiment Rate, together with the classification of its strength, and its verification, constitutes the contribution of this study.

Citations

Author (1)