Behrouz Boochani, the Kurdish-Iranian author of No Friends But the Mountains, has been using social media platforms such as Twitter to speak up against social injustice and human rights abuse against immigrants across the globe. This study proposes an artificial intelligence lifecycle for opinion mining of the dominant sentiments, topics, and emotions in Boochani's social media activism. Sentiment analysis (Vader and Textblob), topic modelling (LDA and NMF), and emotion detection are performed to extract hidden sentiments, topics, and emotions from the data that is collected from his tweets from 2017-2021. The results show Vader performs better than Textblob. LDA is considered the best algorithm. It extracted seven main topics as suicide, translator of book, publication of book, human rights, political, immigration, and detention. Finally, the main emotion detected from the tweets is sadness.
TopIntroduction
Over the past two decades, social media has played a key role in progressive social, political, and cultural changes in the world. As a platform that allows the creation and exchange of User Generated Content, social media is now widely used by human rights activists to report on discriminations and coordinate individual and collective campaigns and actions. One prime example is the literary writer, journalist, and human rights defender, Behrouz Boochani. Aa an asylum-seeker from the Middle East, who was detained by Australia in Manus Island for six years from 2013 until its closure in 2017, Boochani used smartphone and WhatsApp text-messaging application to document and share his experience as a refugee in what became an award-winning book, No Friend But the Mountains (2018). Since fleeing Australia's detention island, Boochani has actively used social media platforms such as Twitter to distribute information on a vast number of topics related to human rights and social justice. His tweets have been followed closely by those monitoring developments in matters related to refugees and life in the Middle East. In this study, we argue that Boochani’s activism in Twitter is a perfect example of how social media has expanded access to evidence of human rights abuses beyond that which is presented by the mainstream media and non-government organizations (NGOs). In this regard, a content analysis of his tweets would be helpful in shedding light on matters that are often overlooked or shrouded by mainstream media.
This study aims to use Opinion Mining techniques to identify the dominant topics in Boochani’s social media activism. The data for this study is collected from his tweets from the beginning of 2017 to the end of 2021. To achieve the main objective of the study, sentiment analysis techniques of TextBlob and VADER rule-based approach are used to study the underlying sentiments. Moreover, Topic Modelling is performed to extract hidden topics by using Latent Dirichlet Allocation (LDA) and Non-negative Matrix Factorization (NMF). The inputs are then processed based on grammatical rules, linguistic habits, and standard algorithms to produce computer-based natural language. Natural language processing (NLP) uses machine learning (ML) systems to ingest and learn words and syntax (Al Mamun et al., 2022; Asri et al., 2022; Fasha et al., 2022; Keikhosrokiani & Asl, 2022; Malik et al., 2021; Paremeswaran et al., 2022; Sofian et al., 2022). Lastly, the dominant emotion is detected from such tweets using LSTM which helps to detect the user’s emotion.
This chapter is organized as follows: It begins with an introduction and a statement of the overall aim of the study. Next is Literature Review that explores sentiment analysis, topic modelling, Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA), and Non-Negative Matrix Factorization (NMF), Probabilistic Latent Semantic Analysis (PLSA), deep learning and emotion detection including Recurrent neural network (RNN), Long Short-Term Memory Units (LSTM), and Recurrent Neural Network- Long Short Term Memory network (RNN-LSTM) Model. After literature review, materials and methodology are presented. Next, results and findings are discussed, and finally, the study is ended with a conclusion that includes recommendations for future studies.