Large language model enhanced public opinion monitoring for China’s portable energy market: A data driven analysis of social media comments

Portable power banks, as a cornerstone of China’s mobile and shared energy market, are closely integrated into daily life and work. The recent, ongoing exposure of safety incidents has led to stricter control measures in specific industries and regions, and even to the revision of relevant administrative management standards, posing a significant challenge to an industry with over a decade of rapid growth. The expectations of the public and government administrative management for product safety, reliability, and service quality have risen rapidly. Therefore, real-time public opinion monitoring is vital for supporting technological innovation, safety governance, and the sustainable development of the industry. However, traditional opinion mining methods struggle to interpret short, noisy, and user-generated data and often lack generalizable semantic understanding. This study therefore develops an LLM-driven framework that first clusters LLM-augmented multi-view representations via graph-based consensus spectral clustering to discover latent topics, followed by multi-label topic assignment and aspect-based sentiment analysis through multi-LLM consensus voting. A total of 47,023 public comments were collected from China’s leading social media platforms, yielding 12 key discussion topics covering electrical safety, charging performance, device durability, shared rental experiences, and emerging magnetic attachment design. The results show that there are strong negative views on electrical safety (86.1%) and the convenience of shared rental services (78.4%), while the views on magnetic connection innovation are mainly positive (76.6%). The topic-event correlation analysis further indicates that opinion fluctuations are closely related to pricing disputes and security-related events. This study demonstrates the effectiveness of LLM-enhanced semantic modeling in public opinion monitoring in the consumer electronics energy field. The results offer reference value for emerging industry market norms under discussion, while the proposed framework can further support the sustainability of portable energy market development by guiding product optimization, risk early warning, and data-informed regulatory strategies.

Authors

Publication Details

Journal
PLoS ONE
Published
2026-09-16
DOI
https://doi.org/10.1371/journal.pone.0357018
Primary Topic
Sentiment Analysis and Opinion Mining
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Large language model enhanced public opinion monitoring for China’s portable energy market: A data driven analysis of social media comments

Jixian Zhou, Jun Wang, Tao Shu, Wei Liao et al.
PLoS ONE
Sentiment Analysis and Opinion Mining
article

Large language model enhanced public opinion monitoring for China’s portable energy market: A data driven analysis of social media comments

Jixian Zhou, Jun Wang, Tao Shu, Wei Liao, Yun Zhang, Weikang Xie, Xiao Li
article en

Abstract

Portable power banks, as a cornerstone of China’s mobile and shared energy market, are closely integrated into daily life and work. The recent, ongoing exposure of safety incidents has led to stricter control measures in specific industries and regions, and even to the revision of relevant administrative management standards, posing a significant challenge to an industry with over a decade of rapid growth. The expectations of the public and government administrative management for product safety, reliability, and service quality have risen rapidly. Therefore, real-time public opinion monitoring is vital for supporting technological innovation, safety governance, and the sustainable development of the industry. However, traditional opinion mining methods struggle to interpret short, noisy, and user-generated data and often lack generalizable semantic understanding. This study therefore develops an LLM-driven framework that first clusters LLM-augmented multi-view representations via graph-based consensus spectral clustering to discover latent topics, followed by multi-label topic assignment and aspect-based sentiment analysis through multi-LLM consensus voting. A total of 47,023 public comments were collected from China’s leading social media platforms, yielding 12 key discussion topics covering electrical safety, charging performance, device durability, shared rental experiences, and emerging magnetic attachment design. The results show that there are strong negative views on electrical safety (86.1%) and the convenience of shared rental services (78.4%), while the views on magnetic connection innovation are mainly positive (76.6%). The topic-event correlation analysis further indicates that opinion fluctuations are closely related to pricing disputes and security-related events. This study demonstrates the effectiveness of LLM-enhanced semantic modeling in public opinion monitoring in the consumer electronics energy field. The results offer reference value for emerging industry market norms under discussion, while the proposed framework can further support the sustainability of portable energy market development by guiding product optimization, risk early warning, and data-informed regulatory strategies.

PLoS ONEVol. 21(9)
Industry, innovation and infrastructure
Openalex Percentile: Top 8%
Sentiment Analysis and Opinion Mining
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.