Efficient Complex Aggregate Queries with Accuracy Guarantee Based on Execution Cost Model over Knowledge Graphs

Knowledge graphs (KGs) have gained prominence for representing real-world facts, with queries of KGs being crucial their application. Aggregate queries, as one the most important parts KG (e.g., “ What is average price cars produced in Germany?”), can provide users valuable statistical insights. An efficient solution aggregate approximate semantic-aware sampling (AQS). This balances query time and result accuracy by estimating an based on random samples collected from a KG, ensuring that relative error bounded predefined error. However, AQS tailored simple exhibits varying performance complex queries. because usually consists multiple each sub-query influences overall processing quality. Setting large bound yields quick results but lower quality, while aiming high-quality demands smaller sub-query, leading to longer time. Hence, devising effective methods executing has emerged significant research challenge within contemporary querying. To tackle this challenge, we first introduced execution cost model original (i.e., supporting queries) founded Taylor’s theorem. aids identifying initial parameters play pivotal role efficiency efficacy AQS. Subsequently, conducted in-depth exploration intrinsic relationship bounds between its constituent sub-queries), then formalized given constraints all sub-queries. Harnessing multi-objective optimization genetic algorithm, refined sub-queries moderate values, achieve balance query. extensive experimental study datasets demonstrated our solution’s superiority effectiveness efficiency.