Abstract
Recently, many users frequently share their opinions on diverse issues using various social media. Therefore, many governments have attempted to establish or improve national policies according to the public opinions captured from the various social media. In this paper, we indicate several limitations of traditional approaches for analyzing public opinions about science and technology and provide an alternative methodology to overcome the limitations. First of all, we distinguish science and technology analysis phase and social issue analysis phase to reflect the fact that public opinion can be formed only when a certain science and technology is applied to a specific social issue. Next, we apply a start list and a stop list successively to acquire clarified and interesting results. Finally, to identify most appropriate documents fitting to a given subject, we develop a new concept of logical filter that consists of not only mere keywords but also a logical relationship among keywords. This study then analyzes the possibilities for the practical use of the proposed methodology thorough its application to discovering core issues and public opinions from 1,700,886 documents comprising SNS, blog, news, and discussion.
Introduction
Traditional text mining often struggles to capture the nuance of public sentiment regarding science and technology (S&T). Standard approaches frequently suffer from contextual blindness, where technology is analyzed in a vacuum rather than at its intersection with social issues. Furthermore, simple keyword searches often return significant noise, and analysts struggle with the "refinement dilemma"—where over-filtering leads to cliché results and under-filtering leads to a chaotic mess of irrelevant data.
This research proposes that public opinion only truly forms when a specific technology intersects with a social issue (e.g., drones intersecting with privacy concerns). By distinguishing between the technology analysis and social issue analysis phases, we can extract higher-quality insights from massive datasets.
Proposed Methodology
The methodology is structured into a four-phase framework designed to move from broad data collection to specific, actionable sentiment analysis.
- Phase 1: S&T Identification via Logical Filters. Instead of simple keyword matching, Boolean logic (AND/OR) defines a technology. A document is only flagged if it contains terms from multiple required thematic groups.
- Phase 2: Social Issue Extraction. Using Topic Modeling (such as Latent Dirichlet Allocation), recurring social themes are identified within the filtered S&T documents.
- Phase 3: Issue Tracking. A temporal analysis monitors how interest in specific sub-issues evolves over time.
- Phase 4: Opinion Analysis. The final stage uses Word Clouds and Word Networks to visualize the relationships between terms and public sentiment.
The Logical Filter and Sequential Dictionary
A core novelty of this approach is the Logical Filter. This treats S&T definitions as structured queries rather than flat keyword lists. For example, to identify documents about "Intellectual Unmanned Aerial Vehicles," the system checks for combinations of terms across different categories to ensure the document is contextually relevant.
Additionally, the methodology employs a Sequential Dictionary approach. It first applies a "Start List" to ensure initial relevance, followed by a "Stop List" to remove common noise terms (e.g., "today," "very"). This two-step refinement balances clarity with depth.
Case Study and Results
The methodology was tested on a corpus of 1,700,886 documents from South Korean sources, including Twitter, Naver Blogs, and major news outlets. The system was applied to two case studies: Infectious Disease Control and Drones.
Key findings include:
- Filtering Efficiency: From the 1.7M documents, the Logical Filters identified 3,404 highly relevant documents for Infectious Disease and 450 for Drones, significantly reducing the manual review burden.
- Temporal Accuracy: The system successfully captured the massive spike in public discourse regarding MERS in late May 2015.
- Qualitative Mapping: For drones, the analysis clearly bifurcated public opinion into two distinct clusters: a "Security/Human Rights" cluster (linked to CCTV and privacy) and an "Efficiency/Logistics" cluster (linked to delivery and cost).
Conclusion
By separating technology from social context and utilizing structured Logical Filters, this methodology provides a more accurate reflection of public sentiment than traditional keyword-based mining. While the creation of these filters currently requires expert intervention, future work will focus on automating the generation of logical structures to further streamline the analysis of national S&T policies.
Summary updated