Integrating Sentiment Analysis, Text Mining and Predictive Clustering for Resilient Supply Chain Management in Industry 5.0: A systematic literature review and future research agenda

The post-pandemic transformation of global industry and the emergence of Industry 5.0 demand supply chain management systems that are smarter, more adaptive and human-centric. Data-driven approaches have become strategic instruments for early detection, prediction and mitigation of disruptions, and the integration of sentiment analysis and text mining on unstructured sources (social media, customer reviews, logistics news) is gaining attention because such data reflects stakeholder responses to supply chain dynamics. However, most prior studies treat sentiment analysis, text mining and predictive clustering as siloed approaches, which becomes problematic when a system must predict ripple effects in real time. An integrated approach combining web scraping, natural language processing and unsupervised learning can transform weak digital signals into operational early warnings without relying solely on internal transactional data. This study constructs a systematic literature review of work integrating these methods in supply chain management for 2019-2026, following the PRISMA 2020 protocol. From thousands of documents traced across Scopus, Web of Science, ScienceDirect, Emerald and IEEE Xplore, 56 relevant articles were analysed in depth. The findings indicate that transformer-based approaches (BERT, RoBERTa, IndoBERT) deliver higher domain-specific sentiment accuracy than classical lexicon-based methods (VADER, TextBlob) and handle multilingual data better. The review also reveals limited use of Indonesian-language data, weak integration between sentiment clusters and ripple-effect simulation, and an absence of studies adapted to Indonesian chemical, mechanical and industrial-systems sectors. Accordingly, we recommend a Sentiment-Driven Predictive Supply Chain Resilience (SD-PSCR) conceptual framework designed for accuracy, multilingual contextualisation, decision interpretability and applicability in Indonesia.