Abstract
The research presents a novel method of creating and visualizing semantic graphs from multi-formatted big textual data using Large Language Models (LLMs). The proposed system leverages advanced natural language processing techniques with LLM capabilities to extract key concepts and relationships from text data visualization. The approach constructs comprehensive knowledge graphs and enables interactive investigation through URL redirection. Our approach involves multi-tiered processing: robust text extraction from various formats (PDF, DOCX, TXT), text segmentation into manageable chunks, and LLM-based analysis to identify key concepts and semantic relationships. The system integrates semantic and contextual relationships into a comprehensive Knowledge Graph, with added values, linking URLs to specific pages within the original inputs. The methodology automates knowledge extraction by enhancing the Knowledge Graph with detailed relationships and robust interactive features, representing a holistic approach. The paper discusses the system architecture, implementation details, potential applications, and future research directions and development.
Recommended Citation
Tripathi, Vishnu; Singh, Azad; Nimmagadda, Shastri; and Mani, Neel, "Big data-guided Knowledge Graph and its Visualization from Multi-Text Formats and Value-Added Interpretation" (2024). International Conference on Information Systems 2024 Special Interest Group on Big Data Proceedings. 1.
https://aisel.aisnet.org/sigbd2024/1