Graph-Based Mapping of Legal Entities: A Methodological Framework For Economic Group Detection In Decision Support Systems

The identification of economic groups is essential for asset recovery and risk analysis. Traditional methods are manual, timeconsuming, and limited. This paper presents a graph-based framework to automate economic group detection by transforming corporate data into graph structures. Using Apache Spark, GraphX, and GraphFrames, the approach integrates semantic similarity models (Sentence-BERT) to reveal hidden corporate links. The system detects both formal and de facto relationships, enhancing decision support while reducing reliance on manual expert analysis. Experimental validation shows that our method expands economic group detection beyond existing tools. This research contributes an automated, scalable methodology applicable to financial investigations, regulatory compliance, fraud detection, and corporate auditing.