Application of intelligent optimization algorithm in constructing visual analysis of knowledge graph dataset

Abstract

This research proposes a multi-stage intelligent optimization framework to enhance knowledge graphs’ visual construction and semantic reasoning capabilities by integrating multiple intelligent optimization algorithms into the layout and analysis process. Considering the complexity and evolving structure of knowledge graph data, analyzing knowledge graph data visually to reveal the internal structure and dynamic evolution has become a key issue. Firstly, the particle swarm optimization (PSO) algorithm is applied to reduce feature dimensionality in large-scale datasets, optimize data quality, and select features useful for knowledge graph construction and analysis. Then, the ant colony optimization (ACO) algorithm is adopted to optimize the path relationship between entities, improving the structural integrity of the knowledge graph and the relationship reasoning accuracy. Next, the grey wolf optimizer (GWO) algorithm is utilized to search for semantic associations in large-scale knowledge graphs efficiently, improving knowledge graphs’ reasoning and semantic understanding capabilities. Finally, the firefly algorithm (FA) is used to optimize node distance and path visualization in the graph layout. Three optimal feature subsets $F_1,F_2$ and $F_5$ are found in the PSO algorithm. The classification accuracy is 90%. The FA algorithm is adopted to obtain the optimized coordinate nodes: A (1.5, 1.5), B (2.5, 3.0), C (4.0, 3.0), D (3.5, 4.5), and E (5.0, 5.0).