Social media sentiment analysis for customer purchasing behavior – A systematic literature review

Abstract

Abstract: Social Media Sentiment Analysis is a field of study with a vast number of applications. One important application is analysing customer behaviours using the results of social media sentiment analysis which is a great tool that decision-makers can utilize. There are several studies conducted about this field. This paper presents the results of a systematic literature review conducted on the existing studies which would be beneficial for developers and researchers interested in this field. This is a preliminary SLR in which, research papers published in journals and conferences until 2020 were collected from 7 electrical databases. Initially, 86 studies were found and 5 most relevant studies derived through specific inclusion and exclusion criteria were investigated to analyse the current status of research, approaches and methods used, results, limitations, existing gaps and future recommendations by researchers. The results of this study suggest that hybrid models that combine lexicons and machine learning classification models produce more accurate results in sentiments analysis. Researchers have attempted to conduct sentiment analysis considering various components of social media text data: punctuation, emoji and emoticons, negations, acronyms and slangs etc. Most studies focus on various applications of social media sentiment analysis beneficial for understanding and interacting with customers. Such as identifying how cultural and economical differences, occurrence of various events impact consumer purchasing behaviours, dealing with negative sentiment shifts, segmenting consumers into groups and even predicting sales performance etc. This study makes a significant contribution by providing a comprehensive and up-todate review of the previous attempts made in the selected domain.