Ontology-based recommendation involving consumer product reviews

摘要

The rapid development of internet, stimulates the development of e-commerce sites. Generally, an e-commerce site provides some review pages, which are collections of reviews from the users of the product. The reviews are useful for a prospective buyer, as consideration for purchasing products. In addition, the recommender system has evolved as a tool for prospective buyers in obtaining the desired product. Most of prior works on knowledge-based recommender system area, utilize recommendation techniques that are based on product features. In this paper, we elaborate the product reviews into the recommendation technique, in addition to product features. Many customers are not familiar with the technical features of the product (e.g. smartphones, cars, cameras, notebooks, etc), so that in our proposed framework, preferences of customers (users) is expressed in the form of functional requirements of products. Mapping the functional requirements - product features - product reviews are in ontology, so the recommendation is based on the exploration of semantic relations in this ontology. To evaluate user's perception of our proposed framework, we compare our proposed framework with a recommender system based on product features (without involving the product reviews), using the Technology Acceptance Model (TAM). Based on evaluation results, we conclude that involving product reviews on recommender system able to increase perceived usefulness and perceived ease of use. In addition, recommender system involving product reviews is more acceptable than recommender system that does not involve product reviews.