Enhanced Knowledge Extraction using Cutting-Edge Rough and Soft Graph Techniques
Surendra Nath Bhagat, Premansu Sekhar Rath, Anirban Mitra · 2025
Addressing uncertainties in data is crucial for real-world applications, including social networks, where relationships often lack clear definition. Soft and rough graph theories allow for nuanced modeling of such data by extending traditional graph structures with parameters that account for uncertainty. Social networks exhibit complex, often ambiguous relationships that traditional graphs cannot fully represent, especially in dynamic, imprecise settings. This study aims to explore soft and rough graphs as effective tools for extracting knowledge from social networks by modeling vague or partially known relationships. Our rough graph analysis identified strong, definite communication channel between employees sharing dual expertise and different weaker channels based on single expertise, suggesting potential collaboration points. In soft graph analysis, two subgroups were detected around shared interests in Sports and Music, identifying central and bridging users within the network. Soft and rough graphs enhance knowledge extraction from social networks, enabling insights into communities, influencers, and communication channels that reflect the uncertain and dynamic nature of social interactions, which can also be used in data science to evaluate ambiguous or incomplete datasets, find hidden patterns, and enhance decision-making in fields like network optimization, fraud detection, and recommendation systems. This approach is shown to be beneficial in structuring and understanding intricate social data patterns.