Ontology-based conversational recommender system for recommending laptop

Abstract

Currently there are many recommender systems for various products, but the recommender system still uses questions that refer to product specifications. For example, on laptop, the question is how much hard drive capacity is needed, what type of processor, etc. With the recommender systems previously described, customers who are not familiar with product's specifications confused in choosing product to match with their needs. To solve these problems, a recommender system is required to prioritizes the functional requirements (high level requirement). Previously, we have developed a multi-domain framework for developing a conversational recommender systems (CRS) based on functional requirements. This framework comprises interaction generate method and ontology. The ontology aims to map functional requirements with technical features of the product. In this paper, we operationalize that framework and propose an ontology based on that framework for developing a CRS in a laptop domain. This CRS interacts with an iterative conversation to find out what the customer needs (e.g., customers need a laptop to watch video). This recommender system does the same conversation as a customer with a professional seller. The users involved in this test show that a recommender system that prioritizes functional requirements is more helpful in product selection than the recommender system commonly used in e-commerce.