Multi-step context-size optimization method for discovering hidden relationship networks from knowledge graphs.

Abstract

The invention relates to a multi-step context-size optimization method for discovering hidden relational networks from knowledge graphs, comprising the following steps: Step 1: Extracting core entities; Step 2: Calculating and assigning dynamic link weights; Step 3: Building an optimal inference network; Step 4: Contextual enrichment; Step 5: Linearization and generating aggregated results. The method reduces information overload, optimizes computational costs, and improves accuracy in data mining compared to manual matching methods. It is applicable in complex query solutions, event investigation, and automated customer support.

Images