A Learning Based Framework for Improving Querying on Web Interfaces of Curated Knowledge Bases

摘要

Knowledge Bases (KBs) are widely used as one of the fundamental components in Semantic Web applications\udas they provide facts and relationships that can be automatically understood by machines. Curated knowledge\udbases usually use Resource Description Framework (RDF) as the data representation model. In order to query\udthe RDF-presented knowledge in curated KBs, Web interfaces are built via SPARQL Endpoints. Currently,\udquerying SPARQL Endpoints has the problems like network instability and latency, which affect the query\udefficiency. To address these issues, we propose a client-side caching framework, SPARQL Endpoint Caching\udFramework (SECF), aiming at accelerating the overall querying speed over SPARQL Endpoints. SECF identifies\udthe potential issued queries by leveraging the querying patterns learned from clients’ historical queries and\udprefecthes/caches these queries. In particular, we develop a distance function based on graph edit distance to\udmeasure the similarity of SPARQL queries. We propose a feature modelling method to transform SPARQL\udqueries to vector representation that are fed into machine learning algorithms. A time-aware smoothing-based\udmethod, Modified Simple Exponential Smoothing (MSES), is developed for cache replacement. Extensive\udexperiments performed on real world queries showcase the effectiveness of our approach, which outperforms\udthe state-of-the-art work in terms of the overall querying speed.