Query refinement in recommender system based on product functional requirements

Abstract

One of the advantages of conversational recommender system (CRS) is its ability to guide user in expressing their needs through conversation mechanism. In many cases, customers are often unable to express their needs clearly at the beginning of the interaction. CRS guides customer (user) to clarify his needs through the mechanism of query refinement. Not all customers are familiar with the technical features of product, especially for hightech product that has many features (complex products). Therefore, asking the customer's needs based on functionalities of product that he/she really wants (product functional requirements) is a more natural way for eliciting customer preference. This paper proposes a model of query refinement that is able to guide users to express their needs based on the functional requirements of product. The model was developed by utilizing exploration of semantic relations on ontology. The evaluation result shows that our proposed model is able to narrow down the average result's size significantly within short interactions.