Sentiment analysis on cryptocurrency in a Muslim-majority country: Evidence from Indonesian social media

  • 22 Views
  • 3 Downloads

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

Type of the article: Research Article

Abstract
The rapid growth of cryptocurrency has generated significant debate in Muslim-majority societies, where financial transactions are expected to conform to Islamic legal principles. In Indonesia, the Indonesian Ulema Council (MUI) issued a fatwa in 2021 declaring the use of cryptocurrency as a means of exchange unlawful (haram) on the grounds that it contains elements of excessive uncertainty and harm. Despite this religious prohibition, the government continues to permit cryptocurrency trading as a regulated commodity, creating a complex normative environment that shapes public discourse. This study analyzes the sentiment of Indonesian Muslims toward cryptocurrency by applying Natural Language Processing (NLP) techniques to a corpus of 1,065 Indonesian-language tweets collected from the X (formerly Twitter) platform over six months (October 2024–March 2025). The IndoBERT-base-p1 model, a state-of-the-art transformer-based language model pre-trained on Indonesian text, was employed for three-class sentiment classification (positive, neutral, negative). The results indicate that neutral sentiment is dominant, accounting for 80.8% of the corpus (861 tweets), while positive sentiment represents 12.9% (137 tweets) and negative sentiment 6.4% (67 tweets). The predominance of neutral sentiment suggests that the majority of Indonesian Muslim social media users maintain a cautious and observational stance, shaped by ongoing scholarly debate, limited digital financial literacy, and regulatory ambiguity. The study contributes to the literature by providing the first application of a deep learning NLP model to the analysis of Indonesian Muslim discourse on cryptocurrency, integrating perspectives from Islamic finance and computational linguistics.

Acknowledgment
The substantial financial support for this study is sponsored by the Ministry of Religious Affairs (MoRA) and the Indonesia Endowment Fund for Education (LPDP) of the Ministry of Finance of the Republic of Indonesia.

view full abstract hide full abstract
    • Figure 1. Research data processing workflow
    • Figure 2. Illustration of the BERT fine-tuning procedure
    • Figure 3. Confusion matrix for the IndoBERT model on the test set
    • Table 1. Classification performance metrics of the IndoBERT-base-p1 model
    • Conceptualization
      Nur Rizqi Febriandika, Annisa Nur Faizah, Helmi Imaduddin
    • Data curation
      Nur Rizqi Febriandika
    • Formal Analysis
      Nur Rizqi Febriandika, Annisa Nur Faizah
    • Funding acquisition
      Nur Rizqi Febriandika
    • Methodology
      Nur Rizqi Febriandika, Annisa Nur Faizah, Helmi Imaduddin
    • Project administration
      Nur Rizqi Febriandika
    • Resources
      Nur Rizqi Febriandika
    • Software
      Nur Rizqi Febriandika, Annisa Nur Faizah, Helmi Imaduddin
    • Supervision
      Nur Rizqi Febriandika
    • Validation
      Nur Rizqi Febriandika, Annisa Nur Faizah
    • Writing – original draft
      Nur Rizqi Febriandika
    • Writing – review & editing
      Nur Rizqi Febriandika, Annisa Nur Faizah, Helmi Imaduddin
    • Investigation
      Helmi Imaduddin