Goods recommendation based on retail knowledge in a Neo4j graph database combined with an inference mechanism implemented in jess

Along with the extensive use of ontologies as a well-established means for knowledge representation, there is a pressing need for methods that can transform ontology information into knowledge stored in a Graph Database (GDB) which is considered human-like thinking in terms of objects and their relations. In this paper, we describe a two-layer knowledge graph database: a concept layer and an instance layer. The concept layer is the resulting graph representation transferred from an ontology representation. The instance layer is the instance data associated with concept nodes. In this research, we apply the two-layer approach to a retail business transaction data for business information query and reasoning. The two-layer structure is implemented in Neo4j GDB platform and information query and recommendation is implemented with a Jess reasoning engine. The query and recommendation results are represented and visualized in knowledge graph structures. The performance of the system is evaluated in terms of the time efficiency of answering queries of retail data using the GDB and the novelty of recommendations.