Document Type



This position paper describes an on-going work on a novel recommendation framework for assisting online shoppers in choosing the most desired products, in accordance with requirements input in natural language. Existing feature-based Shopping Guidance Systems fail when the customer lacks domain expertise. This framework enables the customer to use natural language in the query text to retrieve preferred products interactively. In addition, it is intelligent enough to allow a customer to use objective and subjective terms when querying, or even the purpose of purchase, to screen out the expected products.